test: annulus mass ledger definition with synthetic seam/edge checks

Host-only stdlib definition of the per-hop identity dM + Q_exact = C + H with Q_exact from pre-staging owner populations and H = sum over every annulus cell/direction of (pulled - ideal owner input), split into seam (internal) and edge parts. Synthetic tests: ideal pull, collision source, interior-seam and edge perturbations, the double count of staged flux + full H, a late owner snapshot hiding staging, and a wrong pull direction. No solver change, no GPU. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-20 18:32:42 +00:00
113 changed files with 2594 additions and 3259 deletions
@@ -112,9 +112,8 @@ from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
 cfg = LBMConfig(nx=32, ny=24, viscosity=.1)
 body = SimObject(0, geometry=CircleGeometry(16.0, 12.0, 2.15), center=(16.0, 12.0), radius=2.15,
                 surface_actuation=SurfaceActuation(0.7, 0.85, 0.002))  # angle, half-width, speed
-with CudaContext(0) as ctx, FixedPatchStepper(cfg, context=ctx, refinements=[(8, 6, 24, 18)],
-                                              level_blocks=[None, (16, 20)],
-                                              level_objects=[(), (body,)]) as run:
+with CudaContext(0) as ctx, FixedPatchStepper(cfg, (8, 6, 24, 18), context=ctx,
+                                              fine_objects=[body], fine_block=(16, 20)) as run:
    rest = np.asarray(compute_feq_d2q9(1.0, 0.0, 0.0), np.float32)[:, None, None]
    # (9, ny, nx) postcollision populations per level; non-leaf cells are discarded.
    run.initialize(np.broadcast_to(rest, (9, *run.patch.coarse_shape)),
@@ -125,8 +124,8 @@ with CudaContext(0) as ctx, FixedPatchStepper(cfg, context=ctx, refinements=[(8,

 Scope and limits:
 - The examples below use D2Q9, SRT, FP32, double buffer, unshifted, no LES or
-  forcing. Fixed 2:1 hierarchies use level lists of arbitrary depth; two- and
-  three-level examples below use that same constructor.
+  forcing. Fixed 2:1 hierarchies can use level lists of arbitrary depth; the
+  legacy two/three-level forms are shown below.
 - `coarse_boundary="periodic"` (default) or `"slot_reservoir"` (bottom slot, top
  reservoir; patch needs `y0 >= 4`, `y1 <= ny-4`). Periodic mode ignores the wall/slot config fields.
 - Object centers/radii use the coarse **cell-edge** frame (cell `I` spans `[I, I+1)`),
@@ -134,51 +133,50 @@ Scope and limits:
 - Bodies must be stationary circles whose whole boundary stencil lies in the fine
  active region, clear of the coarse/fine transfer cells; violations are rejected
  before any GPU allocation.
- `level_blocks=[None, (by, bx)]` stores the fine level in fixed-size blocks; `by`/`bx` must
+- `fine_block=(by, bx)` stores the fine level in fixed-size blocks; `by`/`bx` must
  divide the WHOLE fine allocation, i.e. `2*(y1-y0)+8` by `2*(x1-x0)+8` including the
  four-cell ghost band. It gives results identical to dense storage and is
  infrastructure for future block AMR, not a speedup.
 - Several regions: pass a sequence of `(x0, y0, x1, y1)` rectangles as
-  `refinements[0]`, with both levels blocked (`level_blocks`). Transfers are built once
+  `bounds`, with both `fine_block` and `base_block`. Transfers are built once
  from their union; adjacent, overlapping and concave unions are accepted when
  their actual dependencies close. There is no fixed rectangle-gap rule.
  The fine level is then one level-wide grid `(2*ny, 2*nx)`:
  `run.patch.fine_shape` gives it, `initialize`/`download` use it, and only the
  grid-aligned tiles intersecting the union plus its four-cell ghost band are stored.
  Block sizes must divide the level-wide grid; rectangle edges need not align
-  with tiles. Order of rectangles does not matter. L1 bodies/Jets (`level_objects[1]`/`level_jets[1]`) work
+  with tiles. Order of rectangles does not matter. `fine_objects`/`jets` work
  as for one patch: each circle (same coarse cell-edge frame) must lie wholly
  inside the union's active region with
  its boundary stencil clear of transfer cells; a jet may not select links of two
  different bodies, even across regions. Wall-touching patches remain unsupported.
-  For nesting, add further `refinements` entries as shown below.
+  For nesting, use the level lists or legacy child forms below.

  ```python
  cfg = LBMConfig(nx=32, ny=16, viscosity=.1)
  rest = np.asarray(compute_feq_d2q9(1.0, 0.0, 0.0), np.float32)[:, None, None]
  with CudaContext(0) as ctx, FixedPatchStepper(
-          cfg, context=ctx, refinements=[((4, 4, 12, 12), (16, 4, 24, 12))],
-          level_blocks=[(4, 4), (4, 4)]) as run:
+          cfg, ((4, 4, 12, 12), (16, 4, 24, 12)), context=ctx,
+          fine_block=(4, 4), base_block=(4, 4)) as run:
      run.initialize(np.broadcast_to(rest, (9, *run.patch.coarse_shape)),
                     np.broadcast_to(rest, (9, *run.patch.fine_shape)))  # (9, 32, 64)
      run.step(32)
      coarse_g, fine_g = run.download()  # coarse: NaN on every covered cell;
                                         # fine: NaN on every non-ACTIVE cell
  ```
- Three-level nested windows: pass two scalar rectangles in `refinements` —
-  L1 in L0 cells, L2 in **L1 window cells** (the parent's fine grid including its
-  4-cell ghost band) — plus all three `level_blocks` entries. L2 must sit inside
-  L1's ACTIVE cells, away from L1's ghost band and transfer footprint (checked
-  before any GPU allocation). `initialize` and `download` take/return three
-  arrays; each level is NaN on its non-ACTIVE (and child-covered) cells. One
-  coarse step = 2 L1 steps = 4 L2 steps. One L1 window only in this scalar
-  chain. Optional stationary circles and Jets live on L2 via `level_objects[2]`
-  and `level_jets[2]`; L1 bodies/Jets retain their L1 meaning and must be empty
-  in this mode. All object and Jet geometry uses the **L0 edge frame**, not the
-  child bounds' L1 window frame. L2 has dx=dt=0.25, so prescribed lattice
-  velocity is unchanged. Each whole body and its BFL stencil must be inside L2
-  ACTIVE cells, outside the transfer footprint. A Jet may select one physical
-  body across multiple blocks, never two bodies.
+- Legacy three-level scalar chain: add `child=(x0, y0, x1, y1)` in
+  **L1 window cells** (the patch's fine grid including its 4-cell ghost band) and
+  `child_block`, with `fine_block` and `base_block`. L2 must sit inside L1's
+  ACTIVE cells, away from L1's ghost band and transfer footprint (checked before
+  any GPU allocation). `initialize` and `download` take/return three arrays; each
+  level is NaN on its non-ACTIVE (and child-covered) cells. One coarse step = 2 L1
+  steps = 4 L2 steps. One L1 patch only. Optional stationary circles and Jets
+  live on L2 via `child_objects` and `child_jets`; `fine_objects`/`jets` retain
+  their L1 meaning and must be empty in this mode. All object and Jet geometry
+  uses the **L0 edge frame**, not the child bounds' L1 window frame. L2 has
+  dx=dt=0.25, so prescribed lattice velocity is unchanged. Each whole body and
+  its BFL stencil must be inside L2 ACTIVE cells, outside the transfer footprint.
+  A Jet may select one physical body across multiple blocks, never two bodies.

  ```python
  from CelerisLab.body.geometry.circle import CircleGeometry
@@ -190,9 +188,9 @@ Scope and limits:
  jet = Jet(Tube((6.5, 8.), 0., 3., .3), Parallel(0.), amplitude=.002)
  rest = np.asarray(compute_feq_d2q9(1.0, 0.0, 0.0), np.float32)[:, None, None]
  with CudaContext(0) as ctx, FixedPatchStepper(
-          cfg, context=ctx, refinements=[(4, 4, 12, 12), (8, 8, 16, 16)],
-          level_blocks=[(4, 4)] * 3, level_objects=[(), (), (body,)],
-          level_jets=[(), (), (jet,)]) as run:
+          cfg, (4, 4, 12, 12), context=ctx, fine_block=(4, 4), base_block=(4, 4),
+          child=(8, 8, 16, 16), child_block=(4, 4),
+          child_objects=[body], child_jets=[jet]) as run:
      shapes = (run.patch.coarse_shape, run.patch.fine_shape, run.child_patch.fine_shape)
      run.initialize(*(np.broadcast_to(rest, (9, *s)) for s in shapes))  # 16², 24², 24²
      run.step(4)
@@ -200,16 +198,16 @@ Scope and limits:
      run.step(4)
      coarse_g, l1_g, l2_g = run.download()  # solids/covered/nonleaves are NaN
  ```
- Level lists (the only constructor): `refinements[l]` describes level `l+1`
-  in allocated level-`l` integer cells; use a rectangle list for a union, or a
-  scalar rectangle for a single window (a scalar child requires a scalar
-  parent). `level_blocks`, `level_objects` and `level_jets` each have one entry
-  per level, including L0. Three or more levels require all-blocked storage and
-  bodies/Jets only on the finest level. Every adjacent pair must preserve
-  inherited ACTIVE coverage, valid donors/receivers, surviving pulls and the
-  upstream transfer footprint. Level `l` takes `2**l` steps per base interval.
-  This describes implemented topology and scheduling, not accuracy or
-  convergence qualification.
+- Level lists: `refinements[l]` describes level `l+1` in allocated level-`l`
+  integer cells; use a rectangle list for a union, or a scalar rectangle for a
+  single window (a scalar child requires a scalar parent). `level_blocks`,
+  `level_objects` and `level_jets` each have one entry per level, including L0.
+  These lists are mutually exclusive with the legacy patch keywords. Three or
+  more levels require all-blocked storage and bodies/Jets only on the finest
+  level. Every adjacent pair must preserve inherited ACTIVE coverage, valid
+  donors/receivers, surviving pulls and the upstream transfer footprint.
+  Level `l` takes `2**l` steps per base interval. This describes implemented
+  topology and scheduling, not accuracy or convergence qualification.
 - Blocked body levels can enable `observe_raw_exchange=True` and call
  `read_raw_exchange(body_id, level=...)` after a completed base interval.
  It reduces unique cut-link observations into wall/aperture/total force,
@@ -3,13 +3,6 @@
 Changes that can make a previously accepted setup or call behave differently.
 Each entry names the commit, the affected entry points and the observable result.

-## Redundant periodic stepper alias removed
-
-`PeriodicPatchStepper` is no longer exported from `lbm.refinement_stepper`.
-Use `FixedPatchStepper` with the same arguments, including explicit
-`level_blocks`. The old name was an alias for that exact class; periodic
-boundaries remain the default. Existing imports of the old name must be updated.
-
 ## 87c3a31: cut-link ownership is validated for every body

 `validate_cut_link_ownership` (`body/preprocess/flag_overlay.py`) now runs for
@@ -40,18 +33,3 @@ or negative amplitude (the aperture test is geometric). Bodies with neither
 `SurfaceActuation` nor aperture links still raise
 `ValueError("Body has no surface actuation")`. The value is unchanged: the latest
 step's wall/aperture exchange, not total nozzle force.
-
-## 7afb5ad: `Simulation` fails closed after a Jet amplitude device-write failure
-
-`Simulation.set_jet_amplitude` previously only documented that a device error
-leaves the per-link velocity column undefined; the instance stayed usable.
-Now any error after input validation (internal stream creation or the upload)
-marks the Simulation failed. Every later device-touching public method,
-including the `stream` property (guard added in the follow-up repair commit on
-`claude/jet-update-consistency`), raises
-`RuntimeError("Simulation failed (...); close() and recreate it")`. The host
-amplitude keeps its previous value. `close()` still works and stays retryable;
-cached host readers and host staging calls are not blocked. Invalid input
-(`ValueError`/`IndexError`) is unchanged: it raises before any change and the
-instance stays usable. `FixedPatchStepper.set_jet_amplitude` already failed the
-instance (status `failed`) and is unchanged.
@@ -4,7 +4,7 @@
 G-to-G explosion reconstruction. The omitted/default `"constant"` selector retains
 the same kernel, launch arguments and device allocations. Projected mode adds two
 compact tables per adjacent pair; recursion, collision, BFL, restriction, buffer
-ownership and synchronized publication are unchanged. `remap_finest` preserves it.
+ownership and synchronized publication are unchanged. `remap_child` preserves it.

 Runtime eligibility is periodic-root, fully blocked D2Q9/SRT/FP32, body/Jet-free,
 without LES; the fixed-patch backend has no forcing interface. Shifted and unshifted
@@ -25,7 +25,7 @@ under the alternate compile branch.
 - No LES, no external forcing (`d_params` force fields stay zero).
 - Shifted Jets stay SRT-only: `body/jets.py` keeps its existing guard. This extension does not
  provide shifted MRT Jets.
- `PeriodicBlockStepper` (uniform blocked stepper) remained SRT-only; it was removed in 8e1feff and its single-level capability, still SRT-only, is `FixedPatchStepper(refinements=[])`.
+- `PeriodicBlockStepper` (uniform blocked stepper) remains SRT-only.

 ## Evidence status
 CPU: `tests/unit/test_refinement_mrt.py` (host guards, source contract). GPU:
@@ -1,6 +1,6 @@
 # Blocked AMR root channel increment

-Opt in with `FixedPatchStepper(..., coarse_boundary="channel", refinements=[bounds], level_blocks=[(by,bx), (fy,fx)])`. Every level must be blocked. The supported boundary is the existing uniform D2Q9/SRT/FP32 double-buffer channel: local Zou-He west inlet, pressure-NEQ east outlet, stationary halfway bounce-back y walls. Uniform and parabolic inlet profiles and shifted/unshifted storage are supported; the existing configuration guards still reject MRT/LES and incompatible storage. This does not change the external DFG MRT reference.
+Opt in with `FixedPatchStepper(..., coarse_boundary="channel", base_block=(by,bx), fine_block=(fy,fx))`, or the existing level-list form. Every level must be blocked. The supported boundary is the existing uniform D2Q9/SRT/FP32 double-buffer channel: local Zou-He west inlet, pressure-NEQ east outlet, stationary halfway bounce-back y walls. Uniform and parabolic inlet profiles and shifted/unshifted storage are supported; the existing configuration guards still reject MRT/LES and incompatible storage. This does not change the external DFG MRT reference.

 `step/block_channel.cu` is a small root-only kernel. Fine kernels, transfer kernels and recursive hop scheduling are unchanged. Global coordinates and the global profile height/velocity are separate from block-local storage strides. Corner ownership and collision phases match `OneStep`:

@@ -1,77 +0,0 @@
-# src coherence stage (from main 914906e)
-
-Owner direction 2026-10-28: unified style over historical compatibility, inside `src`.
-No numerical change is intended; kernel text was moved verbatim (PTX/bit parity is
-pending a GPU run).
-
-## Conventions (defined once, in `lbm.hierarchy.Hierarchy`)
-
- Level `l = 0 .. depth-1` (0 = root). Pair `hierarchy.pairs[l-1]` builds level `l`.
- Array shapes are `(ny, nx)`; rectangles, origins and centers are `(x, y)`.
- Coarse/fine name the two members (parent/child) of ANY pair, not absolute L0/L1:
-  `fine_shape`, `fine_origin`, `prepare_fine_boundary`, `blocked_fine_tables`,
-  `blocked_coarse_tables`.
- Refined levels (l >= 1) of FixedPatchStepper place bodies/Jets in the root EDGE frame
-  (root cell `i` = `[i, i+1)`). Uniform `Simulation` objects AND the static circles of a
-  single-level (depth 1) FixedPatchStepper use the integer-NODE frame. Nothing converts
-  between them. This concerns input placement only: within the EDGE frame, refined
-  levels still project root-edge coordinates to fine-local cell indices with the
-  half-cell shift (`prepare_fine_boundary`: `(v - o)/dx - .5`; hits back via `fine_hits_to_edge_frame`
-  `origin + dx*(hit + .5)`), using `edge_frame`/`fine_origin` below.
- `Hierarchy.edge_frame(l)` -> `((ox, oy), h)`: cell `(i, j)` of level `l`'s allocated
-  array (ghost band and holes included) covers `[o + h*i, o + h*(i+1))`. A scalar
-  FixedPatch2D window carries the half-coarse-cell ghost shift `bounds - HALO/2` in
-  `fine_origin`; a set's level-wide grid has origin 0. Level 1 boundaries are built from
-  the root config (the wall-aware fields live there); deeper ones from the parent frame.
- `FixedPatch2D.fine_level_offset` (was `fine_level_origin`) is an array INDEX offset
-  `(y, x)`, not an `(x, y)` origin.
-
-## FixedPatchStepper migration map
-
-| removed / changed | replacement |
-|---|---|
-| `patch`, `child_patch` | `hierarchy.pairs[0]`, `hierarchy.pairs[1]` |
-| `base_layout`, `fine_layout`, `child_layout` (stepper properties) | `layouts[0]`, `layouts[1]`, `layouts[2]` (`FixedPatchSet2D.fine_layout(block)` is unchanged) |
-| `fine_boundary`, `child_boundary` | `boundaries[1]`, `boundaries[2]` (host geometry; `None` = no bodies/Jets) |
-| `_jets` | `jets[1]` |
-| `_boundaries[l]` (dict from L1, level keys) | `_device_boundaries[l]` (list from L0, `None` at L0 unless single-level static circles) |
-| `_level_boundaries` (private, b4eacf7) | `_device_boundaries` |
-| `_fine_tables`, `_actuated`, `_jet_cache` | `_device_boundaries[1].tables` / `.actuated` / `.jet_cache` |
-| `_remap_cfg` (private) | `_root_cfg` |
-| `refinement_tables.blocked_base_tables` (also via `refinement_stepper`) | `refinement_tables.blocked_coarse_tables` (import from `refinement_tables`; the stepper no longer re-exports) |
-| `_fine_leaf`, `_child_leaf` | `leaf[1]`, `leaf[2]` (copies) |
-| `_middle_kind` | `hierarchy.kinds[1]` |
-| `_boundary_hosts` (private) | `_boundaries` (private list behind `boundaries`) |
-| `remap_child(rects, child_objects=...)` | `remap_finest(rects, objects=...)` |
-| `set_jet_amplitude(i, a)` (implicit level 1) | `set_jet_amplitude(i, a, level=l)` (required) |
-| `read_raw_exchange(b)` / `read_exchange_window(b)` (implicit level 1) | `read_raw_exchange(b, level=l)` / `read_exchange_window(b, level=l)` (required) |
-| `_level_name(l)` "base/fine/child" messages | `"level l"` |
-| `refinement.MIN_PATCH_GAP` | deleted (no gap policy exists) |
-| `params.cuh HaloPlan`, `collide_trt_no_force` | deleted (zero references) |
-
-Kept: `step_count`, constructor keywords (`refinements`, `level_blocks`, `level_objects`,
-`level_jets`, `coarse_boundary`, `explosion`), `Simulation.set_jet_amplitude(i, a)` (one level),
-`RemapTopology.from_config(..., child_objects=, child_jets=)` (internal remap helper; "child" =
-the replaced level relative to its parent).
-
-## Responsibilities
-
- Construction: `_check_levels` (inputs), `_kernel_configs` (per-level kernel configs),
-  `Hierarchy.edge_frame` (geometry frames), `_build_level` / `_build_pair` / `_upload_boundary`
-  (device). Observation reads share `_observed_boundary` + `_summarize`.
- Blocked kernels: `operators/block_collision.cuh block_collide` is the one collision extension
-  point of `BlockStep` and `BlockChannelStep`. MRT shares one path; SRT keeps its two explicit
-  arithmetic variants (level: unclamped, no-force, shifted direct-momentum feq; channel: uniform
-  OneStep arithmetic with omega clamp), selected by `BLOCK_ROOT_BC`.
- `explosion="projected_linear"` stays an optional capability; `"constant"` stays default.
-
-## Not done / open
-
- Two per-level config derivations stay distinct on purpose: `Hierarchy.configs`
-  (`_middle_config`, wall fields reset on every refined level, geometry validation) vs.
-  `FixedPatchStepper._kernel_configs` (reset only on level 1 of a slot_reservoir root;
-  generates each level's config.h). Folding them needs a config.h/PTX diff gate.
- No single-`Simulation` facade over FixedPatchStepper: the two own different state models
-  (uniform node frame, body sync, forcing, checkpoint vs. immutable blocked hierarchy, edge
-  frame, remap-as-new-instance). A facade would only forward calls; no concrete need found.
- Tests, examples and validation callers are migrated separately (pinned runners stay frozen).
@@ -1,393 +0,0 @@
-# Evidence index (draft, R0)
-
-Active-tree evidence at candidate 0f8fdbc30ea640c764b38ee598d8f4d135c81c85, classified for the release cleanup
-(thread C0C78UFKGRM 1793021833.593279). Draft for review: nothing is removed by this file.
-Recovery of any listed file: `git show 0f8fdbc3:<path>` (blob id given; an immutable tag is requested as anchor).
-
-Classes: K = conclusion kept in the active tree; K-code = tested research code kept; D-snapshot = pre-Gitea
-patch/diff transport (code landed in git from ed1a6ee on; not yet verified hunk by hunk); P-pin = generated hash inventories, git-status dumps and manifests;
-R-raw = run outputs, logs, signals, frames, one-shot run scripts and handoff notes. D, P, R leave the active tree in R1.
-No file under src/ or a test reads any of these paths (git grep at 0f8fdbc3); runners under tests/validation
-that read research/j* are handled in R2.
-
-| Class | Files | Bytes |
-|---|---:|---:|
-| K | 31 | 196,506 |
-| K-code | 3 | 24,324 |
-| D-snapshot | 18 | 338,948 |
-| P-pin | 50 | 979,177 |
-| R-raw | 142 | 3,790,554 |
-
-## (root)
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| BASELINE_JET_AMR.txt | R-raw | 5,244 | 62aa2a66dc05 |
-
-## output
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| round3_compatibility_summary.md | R-raw | 4,366 | f243318f5591 |
-
-## research/fine_surface_20261014
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 54,507 | 71a4543ead37 |
-| cpu_device_guard_cmd.txt | R-raw | 241 | 06e0f35d92b3 |
-| cpu_device_guard_exit.txt | R-raw | 2 | 573541ac9702 |
-| cpu_device_guard_stdout.txt | R-raw | 98 | 424d584a031c |
-| cpu_existing_runs.json | R-raw | 650 | 3179536407b7 |
-| cpu_final_run.json | R-raw | 386 | 523fadb3fec9 |
-| cpu_final_stdout.txt | R-raw | 99 | 9cfe73da3858 |
-| cpu_lifecycle_run.json | R-raw | 387 | 5b403b9e2f30 |
-| cpu_lifecycle_stdout.txt | R-raw | 112 | 3c5c49b285fa |
-| cpu_maps_stdout.txt | R-raw | 95 | 71dd87f9faf9 |
-| cpu_preallocation_run.json | R-raw | 337 | fa6e96d98cdd |
-| cpu_preallocation_stdout.txt | R-raw | 98 | 5d643576f251 |
-| cpu_shared_kernel_stdout.txt | R-raw | 98 | 8a0c046bb9b1 |
-| cpu_surface_command.txt | R-raw | 113 | 2a5ff31e8d06 |
-| cpu_surface_exit.txt | R-raw | 2 | 573541ac9702 |
-| cpu_surface_projection_command.txt | R-raw | 171 | 70b7f5c62c1a |
-| cpu_surface_projection_exit.txt | R-raw | 2 | 573541ac9702 |
-| cpu_surface_projection_stdout.txt | R-raw | 113 | efd37a149cc1 |
-| cpu_surface_stdout.txt | R-raw | 99 | 9cfe73da3858 |
-| gpu_preparation_run1.py | R-raw | 14,043 | ee586d035b7a |
-| gpu_run.json | R-raw | 2,943 | 15b8ba70d528 |
-| gpu_stdout.txt | R-raw | 5,424 | c5216e0267f3 |
-| gpu_test_run1.py | R-raw | 15,479 | f30376abc97f |
-| incremental.diff | D-snapshot | 53,853 | 318d92e7d177 |
-| manifest.json | P-pin | 8,411 | 8281338416af |
-| report.txt | K | 8,373 | 6b9e5123e414 |
-
-## research/fine_surface_joint_20261014
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline_00.json | P-pin | 21,302 | 4ae101ebd02b |
-| baseline_01.json | P-pin | 26,191 | 548cf7b3b494 |
-| baseline_02.json | P-pin | 22,440 | e0421e7a4052 |
-| baseline_03.json | P-pin | 8,584 | da99b45f32a4 |
-| blow_boundary.csv | R-raw | 12,536 | e13363b69457 |
-| blow_collisions.csv | R-raw | 19,145 | 5d5b395486bb |
-| blow_ledger.csv | R-raw | 15,977 | b43bbfd0b8d2 |
-| blow_propagation.csv | R-raw | 9,714 | f06d80d296ca |
-| blow_signals.npz | R-raw | 221,114 | 528adf75baa2 |
-| blow_summary.json | R-raw | 4,113 | 32285ac15be3 |
-| command.txt | R-raw | 232 | 510ecb2925dd |
-| criteria.txt | K | 7,480 | dd36e76a3be1 |
-| frozen.json | R-raw | 11,972 | b5a3324e8686 |
-| gpu_run.json | R-raw | 1,955 | 67ac4785f7e4 |
-| incremental.diff | D-snapshot | 625 | e03c132d22df |
-| integrity.json | P-pin | 826 | a8b6e4be88d2 |
-| manifest.json | P-pin | 4,388 | d2af50260c39 |
-| regions.json | R-raw | 27,525 | 7bba26904221 |
-| report.txt | K | 8,207 | d7571e068ad3 |
-| result.json | R-raw | 11,442 | 9580c2686e70 |
-| review.txt | K | 2,117 | bbe1d8a83d47 |
-| runner.py | R-raw | 5,847 | 812973acc000 |
-| stdout.txt | R-raw | 19,940 | c9218e2e48fc |
-| suction_boundary.csv | R-raw | 12,505 | b870be05e6c0 |
-| suction_collisions.csv | R-raw | 19,173 | 2de48c0a84ab |
-| suction_ledger.csv | R-raw | 16,208 | 92cf1f44911e |
-| suction_propagation.csv | R-raw | 9,720 | f017bcc3ca6f |
-| suction_signals.npz | R-raw | 220,058 | 66d06cf2d909 |
-| suction_summary.json | R-raw | 4,126 | d70a1d371950 |
-| validate.py | R-raw | 19,576 | 37441febd168 |
-| zero_boundary.csv | R-raw | 11,491 | 65c92e48c974 |
-| zero_collisions.csv | R-raw | 18,915 | 027cb8ff6b23 |
-| zero_ledger.csv | R-raw | 16,051 | 34a99bb89dd7 |
-| zero_signals.npz | R-raw | 127,958 | a966f093a9cd |
-| zero_summary.json | R-raw | 1,508 | 6e033f991438 |
-
-## research/fixed_patch_jet_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 69,721 | 91ad9c4d1e88 |
-| before.json | R-raw | 690 | 06e1b6278a29 |
-| fixed_patch_jet.py | R-raw | 9,646 | 6fbdcd9a076b |
-| incremental.patch | D-snapshot | 9,891 | b43129d82c69 |
-| manifest.json | P-pin | 1,967 | 08daa0248f60 |
-| notes.txt | K | 2,214 | 50070fe54efb |
-| results.json | R-raw | 1,100 | d959a6a1cc2c |
-| stdout.txt | R-raw | 1,272 | eccdc0f6b16a |
-
-## research/fixed_patch_jet_closure_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 72,692 | 8a157af82ecf |
-| cpu_results.json | R-raw | 628 | 8e4a1c090c93 |
-| cpu_stdout.txt | R-raw | 590 | d800b783fe47 |
-| incremental.patch | D-snapshot | 3,490 | a9a486c68afd |
-| layout_uncropped.png | R-raw | 51,785 | 0cf1cfa5e99d |
-| manifest.json | P-pin | 1,849 | 31cf0c739241 |
-| notes.txt | K | 3,023 | 0f7e7bdfd0e6 |
-
-## research/fixed_patch_jet_records_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 71,228 | a8e967f76e08 |
-| before.json | R-raw | 690 | ebd6f4251fc2 |
-| fixed_patch_jet.py | R-raw | 10,650 | f6b299885da0 |
-| incremental.patch | D-snapshot | 5,539 | c7f4d0c02ccc |
-| manifest.json | P-pin | 2,057 | 1d616c2541df |
-| notes.txt | K | 2,449 | b58cec14d359 |
-| results.json | R-raw | 1,136 | 788af64a6ebd |
-| stdout.txt | R-raw | 20,750 | a765b06deefa |
-
-## research/j1-2026-10-12
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| REVIEW.md | K | 7,667 | 8f469daef054 |
-| baseline-status.txt | P-pin | 2,601 | 64ca05ab952e |
-| baseline.json | P-pin | 23,052 | ab5852bcb0ef |
-| incremental.diff | D-snapshot | 29,912 | dd9c4e688888 |
-| results.json | R-raw | 68,294 | 56547aa8a678 |
-
-## research/j2-slot-2026-10-12
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| REVIEW.md | K | 12,930 | 69fa4d3bccd3 |
-| baseline-status.txt | P-pin | 2,825 | e711afb0016c |
-| baseline.json | P-pin | 24,660 | 286e023da94f |
-| incremental.diff | D-snapshot | 25,231 | 649f3b5d168d |
-| results.json | R-raw | 477,249 | b4272e25179d |
-
-## research/j3-coverage-2026-10-12
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| REVIEW.md | K | 11,409 | 5887cf73dce2 |
-| baseline-status.txt | P-pin | 3,055 | 78e88356caa3 |
-| baseline.json | P-pin | 30,407 | 095972547a7f |
-| incremental.diff | D-snapshot | 20,114 | f0d502dc5c5b |
-| negative-budget-20744.h5 | R-raw | 88,017 | 693b133be84d |
-| negative-end-20760.h5 | R-raw | 88,174 | 429270d263d4 |
-| restart-21280.h5 | R-raw | 88,062 | 74900affae38 |
-| restart-reference-21296.h5 | R-raw | 88,154 | 7510ee97885b |
-| results.json | R-raw | 296,314 | 7806b10e65cb |
-| verification.json | K | 1,468 | 78a07c47aea4 |
-
-## research/j3-lag-2026-10-12
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| REVIEW.md | K | 11,732 | 34fe5dbb2cae |
-| baseline-status.txt | P-pin | 2,968 | 6ae6094ea9d7 |
-| baseline.json | P-pin | 26,505 | 5c289d14ff30 |
-| end-21280.h5 | R-raw | 92,366 | f30b209179ab |
-| incremental.diff | D-snapshot | 15,790 | 5c6c485e1f75 |
-| results.json | R-raw | 21,718 | 2bf58939129c |
-| start-20760.h5 | R-raw | 92,247 | ae0abd05190c |
-
-## research/j3-offline-2026-10-12
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| REVIEW.md | K | 6,271 | a76e1b486bd2 |
-| lags.csv | R-raw | 1,104 | dd3136dd2dc2 |
-| localization.png | R-raw | 136,013 | 74b56f3e814c |
-| regions.csv | R-raw | 2,126 | 5d361e6872af |
-| results.json | R-raw | 8,852 | bf77b0f534d4 |
-
-## research/j3-reservoir-2026-10-12
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| REVIEW.md | K | 13,563 | a74b67a64e0d |
-| baseline-status.txt | P-pin | 2,896 | a318e077dfae |
-| baseline.json | P-pin | 25,571 | 060502ea7dd8 |
-| incremental.diff | D-snapshot | 22,898 | 8e2c79a86c98 |
-| results.json | R-raw | 182,747 | 7fd5e51a1785 |
-
-## research/refinement_example_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 8,286 | 591478100da8 |
-| before.json | R-raw | 818 | 7ab0962775b1 |
-| incremental.patch | D-snapshot | 8,682 | 7ad8719d0167 |
-| manifest.json | P-pin | 1,798 | bf1eee8a7cf8 |
-| notes.txt | K | 1,185 | 7ff019ad023d |
-| results.json | R-raw | 1,118 | 94697b0988b9 |
-| stdout.txt | R-raw | 325 | 501446f557f4 |
-
-## research/refinement_followup_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| README.txt | K | 4,887 | af65e889f20e |
-| baseline.json | P-pin | 5,343 | 7c60ad8ad692 |
-| cleanup.json | R-raw | 1,253 | 15cf342fba2e |
-| cpu_results.json | R-raw | 1,056 | 1166dffd2a09 |
-| cpu_results_attempt1.json | R-raw | 642 | d139d5508105 |
-| cpu_stdout.txt | R-raw | 499 | d1b33ea20563 |
-| cpu_stdout_attempt1.txt | R-raw | 879 | 6a9fe7fce39a |
-| gpu_before.json | R-raw | 589 | 46762a0ab207 |
-| gpu_before_no_launch.json | R-raw | 589 | 074982d1b2d4 |
-| gpu_rectangle_before.json | R-raw | 589 | 831e3793339f |
-| gpu_rectangle_before_no_launch_1791886104981638567.json | R-raw | 589 | c33885d50bfd |
-| gpu_rectangle_results.json | R-raw | 1,003 | 9c6da219ac6c |
-| gpu_rectangle_stdout.txt | R-raw | 332 | 9185bf303954 |
-| gpu_results.json | R-raw | 1,189 | 53e4f1c44de1 |
-| gpu_seed1_before.json | R-raw | 589 | c77ea3485f84 |
-| gpu_seed1_results.json | R-raw | 996 | b718e8af84e4 |
-| gpu_seed1_stdout.txt | R-raw | 306 | 4d7de0897137 |
-| gpu_stdout.txt | R-raw | 68 | b0956fe3cdb0 |
-| incremental.patch | D-snapshot | 16,836 | 9df25d8c177a |
-| manifest.json | P-pin | 4,410 | 177704d6f1ae |
-| run_gpu.py | R-raw | 5,237 | b5af3f0d4f79 |
-
-## research/refinement_prepare_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| README.txt | K | 8,939 | 8e210c28c7c5 |
-| baseline.json | P-pin | 46,348 | d5f257216b8d |
-| incremental.patch | D-snapshot | 15,751 | 0ddb32550584 |
-| manifest.json | P-pin | 2,113 | 9fd94ccf2c3e |
-| validation.json | R-raw | 3,567 | f057e7cbcd35 |
-| validation.txt | R-raw | 873 | 594daa45c58c |
-
-## research/refinement_slot_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 67,717 | d498081e7c50 |
-| before.json | R-raw | 690 | 430e76b3aa8a |
-| cpu_results.json | R-raw | 1,421 | 4f6bc4d9428f |
-| incremental.patch | D-snapshot | 18,676 | 67a3e8cc9e68 |
-| manifest.json | P-pin | 2,270 | 6515458290f4 |
-| notes.txt | K | 2,612 | 08a713ba4b37 |
-| results.json | R-raw | 1,160 | 03e683027ba8 |
-| stdout.txt | R-raw | 643 | d57b91769826 |
-
-## research/refinement_slot_followup_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 68,536 | 89a26ce79438 |
-| before.json | R-raw | 690 | e707df192ff2 |
-| incremental.patch | D-snapshot | 495 | c6f6fa11237d |
-| manifest.json | P-pin | 1,920 | 4840243a33aa |
-| notes.txt | K | 1,549 | 4b77583ca116 |
-| planar_velocity.cuh | R-raw | 1,911 | 87db7466c843 |
-| planar_velocity.sha256 | P-pin | 251 | 6694c4a1c1f0 |
-| results.json | R-raw | 1,050 | 6ef6eda854c6 |
-| stdout.txt | R-raw | 1,160 | 461860351c0c |
-
-## research/refinement_step_20261013
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| README.txt | K | 10,671 | 88d6f24743f5 |
-| baseline.json | P-pin | 40,620 | 20f61559dd38 |
-| cpu_validation.json | R-raw | 1,231 | 7999fd30018c |
-| gpu_failure_validation.json | R-raw | 868 | 5f6db65ee4f9 |
-| gpu_failure_validation.txt | R-raw | 248 | d9a570a1e9e9 |
-| gpu_validation.json | R-raw | 672 | 2028f59e1055 |
-| gpu_validation.txt | R-raw | 632 | 9699918ee127 |
-| incremental.patch | D-snapshot | 26,516 | d95f9347633d |
-| manifest.json | P-pin | 2,075 | 32a7f1a7309e |
-
-## research/repair-2026-10-10
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| TRON_HANDOFF.md | R-raw | 25,208 | a339786c7944 |
-| baseline-hashes.json | P-pin | 16,020 | 9b4e24cb1d10 |
-| baseline-status.txt | P-pin | 429 | 515be249b638 |
-| checkpoint-domain-2026-10-12.diff | D-snapshot | 14,443 | 035ed630dda1 |
-| classic-results.json | R-raw | 244,729 | 7297aa0d5bca |
-| classic-signals.npz | R-raw | 630,332 | 8cf6d1218aa0 |
-| final-gpu.log | R-raw | 191 | a0eec1a8ea31 |
-| final-unit.log | R-raw | 188 | 94efba2f1d42 |
-| provenance.json | P-pin | 1,020 | 25e22b4f0696 |
-| repair-architecture.md | K | 9,498 | 12051c3e8571 |
-| repair-numerical.md | K | 10,769 | 0c5757fe5f36 |
-| repair-state.md | K | 8,481 | 3c7f2d5b687a |
-| simplification-checks.log | R-raw | 1,045 | fa9d464be740 |
-
-## research/shear_refinement
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| README.md | K-code | 4,142 | 67bdd8ed622b |
-| boundary_shear_tagging.py | K-code | 13,859 | b5415a1c4ea1 |
-| example.py | K-code | 6,323 | 66295d3f0273 |
-
-## research/surface_actuation_20261014
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| baseline.json | P-pin | 54,139 | 0ad06888dbeb |
-| cpu_checkpoint_run.json | R-raw | 400 | 39681b60ff15 |
-| cpu_checkpoint_stdout.txt | R-raw | 114 | 929e4b8d4f46 |
-| cpu_initial.txt | R-raw | 620 | 29213c58a679 |
-| cpu_surface_command.txt | R-raw | 118 | 6d8c615a6e03 |
-| cpu_surface_exit.txt | R-raw | 2 | 573541ac9702 |
-| cpu_surface_stdout.txt | R-raw | 99 | ee9237fc26e5 |
-| gpu_run.json | R-raw | 1,698 | dc8f87eafaa9 |
-| gpu_stdout.txt | R-raw | 159 | b262891e52c3 |
-| incremental.diff | D-snapshot | 50,206 | 43090a1a8b11 |
-| manifest.json | P-pin | 7,527 | 35daa01271cb |
-| report.txt | K | 5,078 | 1162a6479f9f |
-
-## research/wall_seam_pulse_20261014
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| analyze.py | R-raw | 10,343 | ed7cf6999a49 |
-| interpretation.txt | K | 3,514 | 90935db2aa99 |
-| results.json | R-raw | 15,137 | 678a0759871a |
-| run.json | R-raw | 1,170 | 53e303fd0d5d |
-| sources.csv | R-raw | 2,313 | 40d3736b8ef1 |
-| stdout.txt | R-raw | 1,179 | 756d8c38e213 |
-
-## tests/validation/multigrid/cpuref_0_3_evidence
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| README.txt | K | 4,133 | 62897bfd9812 |
-| README_grok.md | K | 1,751 | 2d36e05661a8 |
-| baseline.json | P-pin | 39,291 | 188491096547 |
-| manifest.json | P-pin | 3,495 | 10c19f5fe5ca |
-| results_horsea.json | R-raw | 2,281 | 3b163384e376 |
-| stdout_grok.txt | R-raw | 7,143 | 5a0c270ea4ae |
-| stdout_horsea.txt | R-raw | 9,013 | adff8bb8d295 |
-
-## tests/validation/multigrid/local_stress_v1_evidence
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| README.txt | K | 8,239 | 3513fd294223 |
-| TABLE_local_stress_observable_v1.md | K | 2,967 | 6115dcde0f46 |
-| baseline.json | P-pin | 43,991 | ab631ea4f2ad |
-| comparison.json | R-raw | 13,248 | d7d1b36eb247 |
-| manifest.json | P-pin | 4,132 | c6c44d8125cd |
-| results_horsea.json | R-raw | 2,671 | 79bdf001f948 |
-| run_once.py | R-raw | 4,313 | fa9c5239e993 |
-| source_manifest.json | P-pin | 622 | 0516f1860407 |
-| stdout_horsea.txt | R-raw | 17,091 | a20c554693df |
-| stdout_local_stress_observable_v1.txt | R-raw | 12,238 | 692ffe1a05da |
-
-## tests/validation/multigrid/shear_v0_2_evidence
-
-| File | Class | Bytes | Blob |
-|---|---|---:|---|
-| README.txt | K | 7,220 | 772f74ad40ff |
-| baseline.json | P-pin | 41,751 | 70d9c9fa71c0 |
-| comparison.json | R-raw | 7,679 | 0af8736f9a01 |
-| joint_report.md | K | 6,110 | eed871221298 |
-| manifest.json | P-pin | 3,796 | b22d03a1ba9e |
-| results_horsea.json | R-raw | 1,957 | b2cd48a07a08 |
-| run_once.py | R-raw | 3,245 | 0b8d5e21f326 |
-| source_manifest.json | P-pin | 574 | 4d1c742690c7 |
-| stdout_grok.txt | R-raw | 7,950 | 0e80b1aba7f6 |
-| stdout_horsea.txt | R-raw | 11,421 | 3c867cdac197 |
@@ -1,8 +1,8 @@
 # Fixed three-level rectangle unions

-This page describes three-level `FixedPatchStepper` setups built from the
-`refinements` level list, with rectangle unions on either adjacent pair. The
-same constructor supports arbitrary depth; three or more levels require
+This page describes the legacy three-level entry point of `FixedPatchStepper`,
+with rectangle unions on either adjacent pair. The `refinements` level-list
+entry point also supports arbitrary depth; three or more levels require
 all-blocked storage and bodies/Jets only on the finest level. This is not
 general AMR: there is no per-region timestep, moving geometry, automatic
 dynamic refinement or L1 body support in the three-level form.
@@ -12,15 +12,15 @@ No CUDA kernel changes are involved.

 ## Explicit coordinate frames

-For a single scalar parent patch, pass a scalar rectangle as `refinements[1]`:
-`(x0,y0,x1,y1)` in half-open integer indices of the allocated L1 window,
-including its ghost band.
+Keep the existing scalar `child=(x0,y0,x1,y1)` for a single scalar parent patch.
+Its coordinates retain their original meaning: half-open integer indices in the
+allocated L1 window, including its ghost band.

-A rectangle list as `refinements[1]` accepts a scalar or a set parent and
-interprets every rectangle in the actual allocated L1 array, including ghost
-bands and holes. It never infers region-local coordinates. The physical root
-cell edge corresponding to an L1 index `(x,y)` is
-`parent.fine_origin + (x,y)/2`:
+The new keyword `child_rectangles_in_parent=[...]` is mutually exclusive with
+`child`. It accepts a scalar or a set parent and interprets every rectangle in
+the actual allocated L1 array, including ghost bands and holes. It never infers
+region-local coordinates. The physical root cell edge corresponding to an L1
+index `(x,y)` is `parent.fine_origin + (x,y)/2`:

 - A scalar L1 parent has a shifted window origin.
 - A set L1 parent has origin `(0,0)` and public shape `(2*ny,2*nx)`.
@@ -47,9 +47,9 @@ tile edges.

 ```python
 run = FixedPatchStepper(
-    cfg, context=ctx,
-    refinements=[[(4, 4, 20, 28), (24, 4, 44, 28)], [(16, 16, 32, 40), (56, 16, 80, 40)]],
-    level_blocks=[(4, 4)] * 3,
+    cfg, [(4, 4, 20, 28), (24, 4, 44, 28)], context=ctx,
+    base_block=(4, 4), fine_block=(4, 4), child_block=(4, 4),
+    child_rectangles_in_parent=[(16, 16, 32, 40), (56, 16, 80, 40)],
 )
 ```

@@ -57,11 +57,11 @@ This example uses a 48x32 root. `initialize` and `download` use three Q-major
 dense arrays. Public arrays and masks still scale with the full allocated
 extent; sparse device buffers do not remove that host-memory limit.

-Stationary whole-circle L2 bodies and Jets (`level_objects[2]`, `level_jets[2]`) use root edge
+Stationary whole-circle `child_objects` and `child_jets` use root edge
 coordinates and a single L2 boundary object. Body IDs are physical object-list
 positions, not region IDs or caller `obj_id`. Complete boundary stencils must
 avoid the transfer footprint. Change a child Jet with
-`set_jet_amplitude(index, value, level=2)` (`level` is required); L1 bodies/Jets remain unsupported.
+`set_jet_amplitude(index, value, level=2)`; L1 bodies/Jets remain unsupported.

 ## Verification

@@ -1,26 +0,0 @@
-# Device state lifecycle (uniform Simulation, FixedPatchStepper)
-
-State is grouped by what a change invalidates. This file describes the code as
-it is; it adds no framework. Line references are to candidate 0f8fdbc3.
-
-| Class | Contents | How it changes | Owner rule |
-|---|---|---|---|
-| 1. Module and its globals | Compile-time (visible to the compiler, fixed per module): `config.h` macros incl. `NX`/`NY`, `COLLISION_MODEL`, `BLOCK_ROOT_BC`, `WALL_*`. Runtime per-module globals (not visible to the compiler; written by host code): `__constant__` `d_params`, `d_w`, `d_cx`, `d_cy`, `d_block_channel`, `d_block_wall`; plus bound kernel handles | Globals are written once by the owner right after `load_module`. Uniform grid: `Simulation._recompile` (construction, body-topology sync) loads a new module and re-uploads; `update_runtime_params` rewrites `d_params` (omega, force, n_objects) between runs. Steppers: never rewritten after construction, so they are immutable for the instance's lifetime although they are runtime values. The blocked slot/reservoir root reads its slot from the `d_block_wall` constant (the former dense route used `WALL_*` macros; both were bit-identical in the parity gates). | Every owner loads its own module (`compiler.load_module` = `module_from_file`, no cache). A loaded module is never shared between live owners: globals are per module, so sharing would share mutable `d_params`. |
-| 2. Geometry, layout, tables | flags, curved-link tables, block layouts and halo pairs, transfer (explode/coalesce) tables, block kinds | Immutable for the life of an instance. Uniform grid: rebuilt by body sync. `FixedPatchStepper`: a geometry change is a new instance (`remap_finest`), committed only after a successful transfer. | No in-place table edits. |
-| 3. Controls and dependent boundary columns | Jet amplitude -> per-link `surface_velocity`; body actions (uniform grid) | Between steps, without a table or module rebuild (`set_jet_amplitude`). | Input (index, level, amplitude) is validated before any device write; invalid input changes nothing and the instance stays usable. The host copy commits only after the upload returned. Any error after validation leaves the device column undefined and fails the instance: `FixedPatchStepper` status `failed`, `Simulation._failed`; every later device-touching public method (`Simulation`: `stream`, `set_jet_amplitude`, `initialize`, `sync_bodies`, `run`/`step`, `read_surface_exchange`, `read_body(s)`, `get_macroscopic`, `get_ddf`, `update_runtime_params`, `snapshot`, `restore`, save/load checkpoint) raises; cached host readers and host-side staging (e.g. `get_flags`, `add_body`/`remove_body` before a sync) do not touch the device and stay callable. `close()` still releases resources and stays retryable. |
-| 4. Per-step state | population double buffers and their parity, raw-exchange ledgers, observations | Device-only during `step(n)` / `run(n)`; buffer swaps are host-handle swaps. | Valid on the host only after the single wait at the end of the call. |
-
-Storage routes are separate concerns:
-
- Original uniform grid (`uniform.cu`, `Simulation`): whole-domain `NX`/`NY`.
- Blocked levels (`blocked.cu`; every `FixedPatchStepper` level, including the
-  single level of `refinements=[]`): `NX`/`NY` are one block window including halo.
- Dense refinement levels (`refinement.cu`) were removed: `level_blocks` is
-  required and every refinement level is blocked.
-
-A single-rectangle refinement (`FixedPatch2D`) is stored in its own fine
-window frame (`FixedPatchStepper.__init__`,
-`BlockLayout2D.build(*pair.fine_shape, ...)`). Only a rectangle set
-(`FixedPatchSet2D`) uses the level-wide frame `(2*ny, 2*nx)`, whose
-`fine_layout` requires that full grid to be divisible by the tile
-(`refinement.py:240-252`).
@@ -1,89 +0,0 @@
-# Supported settings: 2D Jet + fixed multilevel refinement (candidate 888c8038)
-
-Every entry names the guard that enforces it (file and function or kernel guard,
-at source revision 888c80381c006e976be2af8a452288729fc17c1b) and, where one
-exists, an existing test that exercises it. A setting not listed here is not
-claimed. Observations at the end are tied to the exact case in which they were
-measured; they are not general bounds.
-
-## Lattice, storage, collision
-
-| Setting | Supported | Guard |
-|---|---|---|
-| Dimension / lattice | 2D D2Q9 only (3D refinement is future work) | `lbm/refinement.py` `FixedPatch2D.from_config`; `#error` guards in `lbm/kernels/blocked.cu` |
-| Precision / streaming | FP32 storage, double buffer | same guards |
-| Collision on refinement levels | SRT or MRT; TRT and LES rejected | `lbm/refinement.py` `FixedPatch2D.from_config` (re-checked for fine boundaries in `prepare_fine_boundary`) |
-| Shifted (ddf_shifting) storage | supported (every level is blocked); one representation on all levels | `lbm/refinement_stepper.py` `FixedPatchStepper._check_storage` |
-| Fine relaxation | 1/tau_fine must lie inside the configured omega guards | `lbm/refinement.py` `FixedPatch2D.from_config` |
-| Single blocked level `FixedPatchStepper(refinements=[], level_blocks=[(by, bx)], level_objects=[circles])` (replaces `PeriodicBlockStepper`, removed in 8e1feff) | doubly periodic, SRT only, no LES/Jets/projection/exchange observation, stationary passive circles; `download()` returns a 1-tuple | `lbm/refinement_stepper.py` `_check_levels`, `FixedPatchStepper.__init__`; `lbm/blocks.py` `static_circle_tables` |
-| Refinement storage | blocked on every level: `level_blocks` is required, one tile `(by, bx)` per level dividing that level; a whole-level tile (`(ny, nx)` at the root, the fine window shape for one rectangle) always divides. Dense and mixed dense/blocked storage were removed (parity of D/D, D/B and B/B recorded at 1573942 and 2469af7) | `lbm/refinement_stepper.py` `_check_levels`, `FixedPatchStepper.__init__` |
-
-Tests: `tests/unit/test_refinement_mrt.py`, `tests/validation/multigrid/test_refinement_mrt_gpu.py`,
-`tests/unit/test_shifted_storage.py`, `tests/unit/test_shifted_block_operators.py`.
-
-## Refinement hierarchy (`FixedPatchStepper(cfg, *, context, refinements, ...)`)
-
-| Setting | Supported | Guard |
-|---|---|---|
-| Depth | at least one refinement level; one entry per level in `level_blocks`/`level_objects`/`level_jets` | `lbm/refinement_stepper.py` `_check_levels` |
-| Single L1 rectangle | two-coarse-cell margin to the domain edge | `lbm/refinement.py` `FixedPatch2D.from_config` |
-| Several L1 rectangles (union) | require `level_blocks[0]` and `level_blocks[1]`; transfers must close (dependency checks, no fixed gap rule) | `lbm/refinement_stepper.py` `_check_levels`; `lbm/refinement.py` `_check_parent_pull_closure` |
-| Nested levels | depth >= 3 must be all blocked; a blocked level requires blocked levels below; a scalar child needs a scalar parent | `lbm/refinement_stepper.py` `_check_levels`; `lbm/hierarchy.py` `build_hierarchy` |
-| Nested transfer support | child only over inherited ACTIVE cells, donors/receivers uncovered ACTIVE, no child over the parent restriction footprint | `lbm/refinement.py` `_nested_kind` |
-| Tile shape | positive `(by, bx)`; the level allocation must be divisible by the tile (no partial tiles yet) | `lbm/refinement_stepper.py` `_check_levels`, `lbm/refinement.py` `FixedPatchSet2D.fine_layout` (shape); `lbm/blocks.py` `BlockLayout2D.build` (divisibility) |
-| Coarse boundary | `periodic`, `slot_reservoir` or `channel` | `lbm/refinement_stepper.py` `FixedPatchStepper.__init__` |
-| `slot_reservoir` | on-site velocity bottom slot, reservoir top, velocity 0, 2 <= slot width < nx, patch rows 4..ny-4 | `lbm/refinement_stepper.py` `FixedPatchStepper.__init__`; `lbm/refinement.py` `prepare_fine_boundary` |
-| `channel` | all levels blocked; Zou-He inlet, pressure-NEQ outlet, stationary bounce-back walls; transfers away from the closure | `lbm/channel.py` `validate_channel_config`, `validate_channel_transfers` |
-| Explosion | `constant` (default); `projected_linear` only periodic, all-blocked SRT, no bodies/Jets, no LES (experimental) | `lbm/refinement_stepper.py` `FixedPatchStepper.__init__` |
-| L2 runtime remap (`remap_finest`) | periodic three-level L2 sets only, no L0/L1 bodies, identical pinned boundary | `lbm/refinement_stepper.py` `FixedPatchStepper.remap_finest` |
-
-Tests: `tests/unit/test_refinement_three_level_multipatch.py`, `tests/validation/multigrid/test_refinement_multipatch.py`,
-`test_refinement_three_level.py`, `test_refinement_three_level_multipatch_gpu.py`, `test_refinement_blocked_base.py`,
-`test_refinement_blocked_fine.py`, `test_channel_root_gpu.py`, `tests/unit/test_channel_root.py`,
-`tests/unit/test_projected_pair_stepper.py`, `test_runtime_child_remap_gpu.py`, `test_pinned_body_remap_gpu.py`.
-
-## Bodies and Jets on refined levels
-
-| Setting | Supported | Guard |
-|---|---|---|
-| Bodies | stationary circles without sensor/force-region roles, entirely inside the patch union, fully resolved, disjoint, stencil away from the two-hop transfer footprint | `lbm/refinement.py` `prepare_fine_boundary` |
-| Level | not on L0; only on the finest level of a nested hierarchy | `lbm/refinement_stepper.py` `_check_levels` |
-| Jet condition | velocity only; `pressure` is reserved and raises NotImplementedError | `body/jets.py` `Jet.__post_init__` |
-| Jet geometry | each Jet selects resolved links of exactly one stationary body; no Jet/Jet or Jet/SurfaceActuation overlap; abs(amplitude) < 1 lattice | `body/jets.py` `Jet.__post_init__`, `jet_link_shapes`, `_surface_links` |
-| Jets with shifted storage | only in the all-blocked stepper with SRT and no LES | `body/jets.py` `_surface_links` |
-| SurfaceActuation | unshifted storage only | `body/objects.py` `SimObject.validate_surface_actuation` |
-| Raw boundary exchange (force) | blocked body levels only; the window ledger requires `observe_raw_exchange` | `lbm/refinement_stepper.py` `FixedPatchStepper.__init__` |
-| Checkpoint with Jets | refused | `simulation.py` `Simulation.save_checkpoint` |
-
-Raw exchange is the cut-link momentum reaction (wall + aperture), not a validated
-nozzle thrust or stress load. Jet amplitude is a lattice velocity, not a flow rate.
-
-Tests: `tests/unit/test_jets.py`, `test_jets_review.py`, `test_shifted_jets.py`, `test_raw_exchange.py`,
-`tests/validation/multigrid/test_refinement_jets.py`, `test_shifted_jets_gpu.py`, `test_raw_exchange_gpu.py`,
-`test_exchange_window.py`, `test_fine_surface_gpu.py`.
-
-## Observations from specific tested cases (not general bounds)
-
- SRT channel root vs uniform OneStep, strict-parity test
-  `tests/validation/multigrid/test_channel_root_gpu.py::ChannelRootGPU::test_uniform_matched_root`:
-  in one of its 18 sub-cases (2 storage modes x 3 profile/clamp/damp settings x 3 blocks; which
-  one is not recorded), 1 of the 4320 compared values differs, max abs 2.33e-10, max rel 1.77e-7
-  (about 2 ULP). The test remains failed; its cause and any general bound are unresolved.
- MRT root parity test
-  `tests/validation/multigrid/test_refinement_mrt_gpu.py::RefinementMRTGPU::test_uniform_matched_mrt_root`
-  (12 sub-cases: 2 storage modes x 2 profiles x 3 blocks, default outlet settings) passed bitwise
-  in the Horsea focused GPU run at K1 7813334 (2026-10-21).
- Periodic Taylor-Green Re 100, measured at 36ffc8b1 (V3 matched pair, N128 depth 2):
-  every per-sample screen passed; the constant explosion's aggregate worst velocity L2,
-  0.004418, exceeds its preregistered 0.002 budget, while projected_linear passed that budget.
- Curved-wall pressure (pilot harness, exact-q analytic annulus, tau 0.8): the coarse swirl r1
-  case exceeded its terminal density-gain cap (1.105e-4 vs 1e-4), so the frozen 9-case matrix
-  is NOT_ESTABLISHED. No validated fix exists. This is a pilot harness, not production code.
- DFG 2D-1 Re 20 (V4, SRT, t* 60, pressure-NEQ outlet and local Zou-He inlet): refined and uniform
-  runs were NOT_STEADY under the frozen steady rule, so no interval verdict was given.
-
-## Not supported in this release
-
-3D, dynamic/automatic AMR (tagging, time-varying hierarchies beyond `remap_finest`),
-moving bodies on refined levels, pressure Jets, Jet checkpoints, LES or TRT on refined
-levels, partial (non-dividing) tiles.
@@ -33,7 +33,7 @@ origin is retained. Physical lower edges are `origin + spacing*(x,y)` on L0,
 `origin + spacing*(parent.fine_origin + (x,y)/4)` on L2.

 Factory options `coarse_boundary` and `objects` must declare `"periodic"` and
-an empty collection; L2 bodies use `objects` as described below.
+an empty collection; L2 bodies use `child_objects` as described below.
 LBMConfig's channel settings do not select this operator's boundary mode.
 Geometry metadata is factory-created;
 manual mutation/forging of metadata is unsupported. Caller certifies the
@@ -77,7 +77,7 @@ PYTHONPATH=src python -m unittest discover -s tests/unit -p 'test_synchronized_r

 ## Explicit runtime continuation

-`replacement = current.remap_finest(rectangles)` returns a new ready
+`replacement = current.remap_child(rectangles)` returns a new ready
 `FixedPatchStepper` at the **same absolute step_count**. It supports only
 periodic three-level runs with no L0/L1 bodies and an existing L2
 `FixedPatchSet2D`; stationary L2 bodies/Jets follow the support contract below.
@@ -99,7 +99,7 @@ resources; an actual old-device synchronization/read failure marks the old
 stepper failed, consistent with its existing failure semantics.

 ```python
-replacement = current.remap_finest([(20, 16, 36, 40)])
+replacement = current.remap_child([(20, 16, 36, 40)])
 # Explicitly choose the returned continuation only after success.
 current.close()
 current = replacement
@@ -119,7 +119,7 @@ Per-population conservation is not an advancement target under collision.

 ### Bounded pinned-body extension

-`replacement = current.remap_finest(rectangles, objects=original_objects)`
+`replacement = current.remap_child(rectangles, child_objects=original_objects)`
 also accepts stationary Circles inside the L2 union. Each body's **physical
 boundary, complete BFL support and current per-link velocity must remain
 identical**. Rectangles are not pinned: they may be split, merged or resized,
@@ -1,9 +1,9 @@
 """Public API demo: fixed 2:1 patch with a stationary, optionally actuated circle.

 From this checkout, after checking the selected physical GPU is idle:
-    PYTHONPATH="$PWD/src" CUDA_VISIBLE_DEVICES=0 python examples/fixed_patch_circle.py
-    ... --speed -0.002 --mode slot_reservoir
-Both levels use blocked storage (--base-block, --fine-block tiles).
+    PYTHONPATH="$PWD/src" CUDA_VISIBLE_DEVICES=0 python examples/fixed_patch_circle.py --fine-block 16 20
+    ... --speed -0.002 --fine-block 16 20 --mode slot_reservoir
+Omit --fine-block for the (equivalent) dense fine storage.

 A short smoke run (default 128 coarse steps): it shows setup, stepping and
 fluid-only diagnostics, not a developed flow, accuracy or long-time stability.
@@ -71,10 +71,8 @@ def main():
    parser.add_argument("--mode", choices=("periodic", "slot_reservoir"), default="periodic")
    parser.add_argument("--speed", type=float, default=.002,
                        help="normal surface velocity, coarse lattice units (+ blows, - sucks, 0 passive)")
-    parser.add_argument("--base-block", type=int, nargs=2, metavar=("BY", "BX"), default=(12, 8),
-                        help="root tile; must divide the 24 x 32 root")
-    parser.add_argument("--fine-block", type=int, nargs=2, metavar=("BY", "BX"), default=(16, 20),
-                        help="fine tile; must divide the whole fine allocation (32 x 40 here)")
+    parser.add_argument("--fine-block", type=int, nargs=2, metavar=("BY", "BX"),
+                        help="blocked fine storage; must divide the whole fine allocation (32 x 40 here)")
    parser.add_argument("--steps", type=int, default=128)
    args = parser.parse_args()
    if args.steps < 0:
@@ -88,20 +86,20 @@ def main():
    body = SimObject(0, geometry=CircleGeometry(*CENTER, RADIUS), center=CENTER, radius=RADIUS,
                     surface_actuation=spec)
    cfg = config(args.mode)
+    fine_block = tuple(args.fine_block) if args.fine_block else None
    # The caller owns the CUDA context; the stepper must be closed first (the
    # nested with-statement does that).
    with CudaContext(0) as context, FixedPatchStepper(
-            cfg, context=context, coarse_boundary=args.mode, refinements=[BOUNDS],
-            level_blocks=[tuple(args.base_block), tuple(args.fine_block)],
-            level_objects=[(), (body,)]) as run:
+            cfg, BOUNDS, context=context, coarse_boundary=args.mode,
+            fine_objects=[body], fine_block=fine_block) as run:
        p = run.patch
        # Rest equilibrium is already postcollision G, so no initial collision is
        # needed. Non-leaf entries (covered, ghost, solid) are discarded by initialize.
        rest = np.asarray(compute_feq_d2q9(1.0, 0.0, 0.0), np.float32)[:, None, None]
        run.initialize(np.broadcast_to(rest, (9, *p.coarse_shape)),
                       np.broadcast_to(rest, (9, *p.fine_shape)))
-        storage = f"blocked root {tuple(args.base_block)} fine {tuple(args.fine_block)}"
-        print(f"mode={args.mode} storage={storage} speed={args.speed:+g} "
+        storage = f"blocked {fine_block}" if fine_block else "dense"
+        print(f"mode={args.mode} fine storage={storage} speed={args.speed:+g} "
              f"(fine allocation {p.fine_shape[0]}x{p.fine_shape[1]}, fine origin {p.fine_origin})")
        masks = leaf_masks(run, args.mode)
        start = leaf_diagnostics(*run.download(), masks)
@@ -31,7 +31,6 @@ from CelerisLab.lbm.refinement import ACTIVE
 from CelerisLab.lbm.refinement_stepper import FixedPatchStepper

 BOUNDS, AMPLITUDE, WIDTH = (16, 4, 32, 28), .02, 12
-LEVEL_BLOCKS = [(12, 8), (8, 8)]  # blocked storage: 48x48 root, 56x40 fine window
 TIMES, MAX_MACH, MAX_SECONDS = (0, 300, 1200), .1, 120


@@ -133,8 +132,7 @@ def main():
                  lattice_model='D2Q9', collision='SRT', streaming='double_buffer',
                  store_precision='FP32', ddf_shifting=False, les_enabled=False)
    report = dict(config=dict(nx=48, ny=48, bounds=BOUNDS, W=WIDTH, A=AMPLITUDE, nu=.1,
-                              rho=1, tau_coarse=.8, tau_fine=1.1, coarse_boundary='slot_reservoir',
-                              level_blocks=LEVEL_BLOCKS),
+                              rho=1, tau_coarse=.8, tau_fine=1.1, coarse_boundary='slot_reservoir'),
                  guards=dict(check_every=50, positive_density=True, finite=True,
                              max_Mach=MAX_MACH, max_seconds=MAX_SECONDS,
                              timing='cooperative checks, not an exact interruption deadline'),
@@ -150,9 +148,8 @@ def main():
                  samples=[], status='running', last_synchronized_coarse_step=0)
    start, run = time.monotonic(), None
    try:
-        with CudaContext(0) as context, FixedPatchStepper(cfg, context=context,
-                coarse_boundary='slot_reservoir', refinements=[BOUNDS],
-                level_blocks=LEVEL_BLOCKS) as run:
+        with CudaContext(0) as context, FixedPatchStepper(cfg, BOUNDS, context=context,
+                coarse_boundary='slot_reservoir') as run:
            p = run.patch
            masks = [np.ones(p.coarse_shape, bool), p.fine_kind == ACTIVE]
            masks[0][[0, -1]] = False
@@ -16,7 +16,7 @@ from CelerisLab.config import load_lbm_config
 from CelerisLab.cuda.context import CudaContext
 from CelerisLab.lbm.equilibrium import compute_feq_d2q9, compute_macro_from_ddf
 from CelerisLab.lbm.refinement import ACTIVE
-from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
+from CelerisLab.lbm.refinement_stepper import PeriodicPatchStepper


 def shear_equilibrium(shape, origin, dx, period):
@@ -48,8 +48,7 @@ def main():
                  nx=16, ny=16, viscosity=.1, velocity=1e-3, rho=1.0,
                  lattice_model="D2Q9", collision="SRT", streaming="double_buffer",
                  store_precision="FP32", ddf_shifting=False, les_enabled=False)
-    with CudaContext(0) as context, FixedPatchStepper(cfg, context=context, refinements=[(4, 4, 12, 12)],
-                                                           level_blocks=[(16, 16), (24, 24)]) as run:
+    with CudaContext(0) as context, PeriodicPatchStepper(cfg, (4, 4, 12, 12), context=context) as run:
        p = run.patch
        coarse = shear_equilibrium(p.coarse_shape, (0, 0), 1.0, cfg.ny)
        fine = shear_equilibrium(p.fine_shape, p.fine_origin, .5, cfg.ny)
@@ -207,30 +207,19 @@ def surface_link_velocity(objects, cfg, flags, fluid_idx, dir, body_id, rx, ry,
                          jets=(), hits=None, *, _shifted_blocked=False):
    """Per-link prescribed surface velocity (n, 2) FP32 and aperture mask.

-    See _surface_links, which also returns the unit-amplitude jet shapes.
-    """
-    return _surface_links(objects, cfg, flags, fluid_idx, dir, body_id, rx, ry, jets, hits,
-                          _shifted_blocked=_shifted_blocked)[:2]
-
-
-def _surface_links(objects, cfg, flags, fluid_idx, dir, body_id, rx, ry,
-                   jets=(), hits=None, *, _shifted_blocked=False):
-    """(velocity, aperture, jet shape, jet id); the last two are None without jets.
-
    Host-only and storage-neutral: rows follow the given physical cut-link
-    order. Returns all None when neither an object carries surface
+    order. Returns (None, None) when neither an object carries surface
    actuation nor a jet is given. Jets need true FP64 ``hits`` (n, 2) and may
    not share a link with each other or with a legacy aperture.
    Shifted storage is privately enabled only by the all-blocked stepper.
    """
    enabled = [obj for obj in objects if obj.surface_actuation is not None]
    if not enabled and not jets:
-        return None, None, None, None
+        return None, None
    from .preprocess.flag_overlay import validate_cut_link_ownership
    validate_cut_link_ownership(flags, fluid_idx, dir, cfg.nq)
    surface_velocity = np.zeros((len(dir), 2), np.float32)
    surface_aperture = np.zeros(len(dir), bool)
-    shape = jet_id = None
    for obj in enabled:
        obj.validate_surface_actuation(cfg)
        selected = body_id == obj.obj_id
@@ -263,4 +252,4 @@ def _surface_links(objects, cfg, flags, fluid_idx, dir, body_id, rx, ry,
            raise ValueError("Jet overlaps a SurfaceActuation aperture")
        surface_velocity[driven] = jet_link_velocity(jets, shape, jet_id)[driven]
        surface_aperture |= driven
-    return surface_velocity, surface_aperture, shape, jet_id
+    return surface_velocity, surface_aperture
@@ -11,7 +11,7 @@ Responsibilities:

 Design::
  Packed ``obs`` layout matches ``generate_config(..., n_objects=count)`` and
-  ``config_obs.h`` macros. Host sizing uses :func:`~CelerisLab.cuda.compiler.obs_layout`.
+  ``config_obs.h`` macros. Host sizing uses :func:`~CelerisLab.cuda.compiler_v2.obs_layout`.

  Pending edits are lightweight: only the "intent" (add / remove) is stored.
  All derived quantities (flags, compact lists, GPU buffers) are built at
@@ -28,7 +28,7 @@ import pycuda.driver as cuda
 from .objects import SimObject
 from .registry import BodyRegistry
 from .preprocess.flag_overlay import merge_flag_masks, validate_cut_link_ownership
-from ..cuda.compiler import obs_layout
+from ..cuda.compiler_v2 import obs_layout
 from ..cuda.obs_layout import ObsLayout
 from ..lbm.descriptors import (
    FLUID, SOLID,
@@ -7,7 +7,7 @@ Two JSON files:
  config_body.json – object list

 LBMConfig / BodyConfig are plain dataclasses; they translate to
-macro dicts consumed by `cuda/compiler.py` for header generation.
+macro dicts consumed by `cuda/compiler_v2.py` for header generation.
 """

 import json
@@ -60,14 +60,14 @@ Python `config.py` 只负责读取和校验，不是配置位置。显式配置

 **`lbm/kernels/config/*.h` 由编译器根据 `LBMConfig` 自动生成，请勿手改。**

-仓库中检入的 `config/*.h` 可能是历史版本。`compiler.generate_config` 会复制 kernel 源码到独立临时构建目录，并在副本中生成配置；不会改写包内头文件或 PTX。返回目录须显式传给 `compile_kernel(build_dir=...)`，编译结果与配置摘要均保存在该目录。`Simulation` 负责目录清理；直接使用编译器的调用者须调用 `cleanup_build_dir()`。跳过生成配置直接编译仓库中的 `uniform.cu` 属不受支持用法。
+仓库中检入的 `config/*.h` 可能是历史版本。`compiler_v2.generate_config` 会复制 kernel 源码到独立临时构建目录，并在副本中生成配置；不会改写包内头文件或 PTX。返回目录须显式传给 `compile_kernel(build_dir=...)`，编译结果与配置摘要均保存在该目录。`Simulation` 负责目录清理；直接使用编译器的调用者须调用 `cleanup_build_dir()`。跳过生成配置直接编译仓库中的 `kernel_v2.cu` 属不受支持用法。

 ```
 config_lbm.json
  ↓  load_lbm_config()
 config.py:LBMConfig       (读取 + 校验)
  ↓  to_macros()
-cuda/compiler.py          (独立构建目录中生成 config/*.h)
+cuda/compiler_v2.py       (独立构建目录中生成 config/*.h)
  ↓
 config_grid.h             → NX, NY, NZ, NT, LATTICE_MODEL；
                            DIM / NQ 由 LATTICE_MODEL 推导写入（与 grid.lattice_model 一致，非 JSON 顶层字段）
@@ -1,4 +1,4 @@
 """CUDA build helpers (compilation itself does not need a live context)."""
-from . import compiler
+from . import compiler_v2 as compiler

 __all__ = ["compiler"]
@@ -1,13 +1,11 @@
-# CelerisLab/cuda/compiler.py
+# CelerisLab/cuda/compiler_v2.py
 """
 Kernel configuration, compilation, and module loading.

-Owners call build_module(cfg, entry, arch=...), then set_params(module, omega, ...).
-build_module runs, in a private directory it removes afterwards:
+Workflow:
  1. build_dir = generate_config(lbm_cfg, n_objects) — isolated source snapshot
-  2. compile_kernel(arch, source=entry, build_dir=build_dir) — nvcc -ptx
+  2. compile_kernel(arch, build_dir=build_dir)       — nvcc -ptx kernel_v2.cu
  3. load_module(ptx_path)              — PyCUDA module from PTX
-Entries: uniform.cu (uniform grid), blocked.cu (every blocked level, single or refined).
 """

 from __future__ import annotations
@@ -18,6 +16,7 @@ import shutil
 import tempfile
 import hashlib
 import json
+from contextvars import ContextVar
 from dataclasses import asdict
 from pathlib import Path
 from typing import Optional
@@ -34,6 +33,9 @@ _KERNEL_DIR = os.path.join(
 )


+_last_build_dir = ContextVar("celeris_last_build_dir", default=None)
+_last_ptx_path = ContextVar("celeris_last_ptx_path", default=None)
+

 def kernel_path(filename: str) -> str:
    return os.path.join(_KERNEL_DIR, filename)
@@ -52,16 +54,14 @@ def _write(path: str, content: str):


 def generate_config(cfg: LBMConfig, n_objects: int = 0,
-                    *, build_root: Optional[str] = None, block_root_bc: int = 0) -> str:
+                    *, build_root: Optional[str] = None) -> str:
    """Snapshot kernels and render headers in a fresh private directory.

-    Returns its path; pass it to compile_kernel(build_dir=...). The caller owns cleanup_build_dir() once the loaded module is no longer
+    Returns its path. Pass it explicitly to compile_kernel for interleaved
+    builds. Legacy generate_config(); compile_kernel() remains context-local.
+    The caller owns cleanup_build_dir() once the loaded module is no longer
    needed for recompilation/debugging. No package files are changed.
-    block_root_bc specializes a blocked root module; fine/dense builds use 0.
    """
-    if (isinstance(block_root_bc, bool) or not isinstance(block_root_bc, int)
-            or block_root_bc not in (0, 1, 2)):
-        raise ValueError("block_root_bc must be an integer mode 0, 1 or 2")
    macros = cfg.to_macros()
    if isinstance(n_objects, bool) or not isinstance(n_objects, int) or n_objects < 0:
        raise ValueError("n_objects must be a nonnegative integer")
@@ -71,19 +71,20 @@ def generate_config(cfg: LBMConfig, n_objects: int = 0,
    try:
        shutil.copytree(_KERNEL_DIR, build_dir, dirs_exist_ok=True,
                        ignore=shutil.ignore_patterns("*.ptx", "*.cubin", "*.tmp"))
-        _render_config(macros, n_objects, os.path.join(build_dir, "config"),
-                       block_root_bc=block_root_bc)
+        _render_config(macros, n_objects, os.path.join(build_dir, "config"))
        digest = hashlib.sha256()
        for path in sorted(Path(build_dir).rglob("*")):
            if path.is_file():
                digest.update(str(path.relative_to(build_dir)).encode())
                digest.update(path.read_bytes())
        metadata = {"source_sha256": digest.hexdigest(), "config": asdict(cfg),
-                    "n_objects": n_objects, "block_root_bc": block_root_bc}
+                    "n_objects": n_objects}
        _write(os.path.join(build_dir, ".celeris-build"), json.dumps(metadata, sort_keys=True))
    except BaseException:
        shutil.rmtree(build_dir)
        raise
+    _last_build_dir.set(build_dir)
+    _last_ptx_path.set(None)
    return build_dir


@@ -95,9 +96,12 @@ def cleanup_build_dir(build_dir: Optional[str]) -> None:
    if path == Path(_KERNEL_DIR).resolve() or not (path / ".celeris-build").is_file():
        raise ValueError("Refusing to remove a directory not owned by the compiler")
    shutil.rmtree(path)
+    if _last_build_dir.get() == str(build_dir):
+        _last_build_dir.set(None)
+        _last_ptx_path.set(None)


-def _render_config(m: dict, n_objects: int, cdir: str, *, block_root_bc: int = 0):
+def _render_config(m: dict, n_objects: int, cdir: str):
    _write(os.path.join(cdir, "config_grid.h"), f"""{_HEADER}\
 // Layer: Global / Grid
 #ifndef CELERIS_CONFIG_GRID_H
@@ -161,9 +165,6 @@ def _render_config(m: dict, n_objects: int, cdir: str, *, block_root_bc: int = 0
 #define STORE_PRECISION  {m['STORE_PRECISION']}
 #define USE_DDF_SHIFTING {m['USE_DDF_SHIFTING']}

-// Internal blocked root mode: 0=periodic/fine, 1=slot/reservoir, 2=channel.
-#define BLOCK_ROOT_BC {block_root_bc}
-
 #define USE_LES {m['USE_LES']}
 #define LES_CS  {m['LES_CS']}
 #define LES_CLOSED_FORM {m['LES_CLOSED_FORM']}
@@ -240,9 +241,13 @@ def _render_config(m: dict, n_objects: int, cdir: str, *, block_root_bc: int = 0
 # ---------------------------------------------------------------------------
 # Compilation
 # ---------------------------------------------------------------------------
-def compile_kernel(arch: str = "sm_70", source: str = "uniform.cu",
-                   output: Optional[str] = None, *, build_dir: str) -> str:
-    """Compile the snapshot in build_dir (from generate_config); never package headers."""
+def compile_kernel(arch: str = "sm_70", source: str = "kernel_v2.cu",
+                   output: Optional[str] = None, *,
+                   build_dir: Optional[str] = None) -> str:
+    """Compile a generated private source snapshot; never use package headers."""
+    # A failed attempt must not make implicit load_module() reuse older code.
+    _last_ptx_path.set(None)
+    build_dir = build_dir or _last_build_dir.get()
    if not build_dir or not (Path(build_dir) / ".celeris-build").is_file():
        raise RuntimeError("Call generate_config() and pass its build_dir before compiling")
    root = Path(build_dir).resolve()
@@ -267,36 +272,23 @@ def compile_kernel(arch: str = "sm_70", source: str = "uniform.cu",
    finally:
        if os.path.exists(tmp):
            os.unlink(tmp)
+    _last_ptx_path.set(str(dst))
    return str(dst)


-def load_module(ptx_path: str):
-    """Load one compiled PTX file as a new PyCUDA module."""
+def load_module(ptx_path: Optional[str] = None):
+    """Load explicit PTX or the last PTX compiled in this Python context."""
+    ptx_path = ptx_path or _last_ptx_path.get()
+    if not ptx_path:
+        raise RuntimeError("No compiled PTX; call compile_kernel() first")
    import pycuda.driver as cuda
    return cuda.module_from_file(ptx_path)


-def build_module(cfg: LBMConfig, source: str, *, arch: str, n_objects: int = 0,
-                 block_root_bc: int = 0):
-    """Generate, compile and load one private module; the build directory is removed.
+# ---------------------------------------------------------------------------
+# Back-compat helpers (used by test scripts)
+# ---------------------------------------------------------------------------

-    The single build path of every owner (Simulation, FixedPatchStepper levels).
-    Each call returns a NEW module: callers must not share it, because its
-    globals (d_params, ...) are mutable per module.
-    """
-    build_dir = generate_config(cfg, n_objects=n_objects, block_root_bc=block_root_bc)
-    try:
-        return load_module(compile_kernel(arch, source=source, build_dir=build_dir))
-    finally:
-        cleanup_build_dir(build_dir)
-
-
-def set_params(module, omega: float, fx: float = 0.0, fy: float = 0.0, fz: float = 0.0,
-               n_objects: int = 0) -> None:
-    """Write the module's __constant__ LBMParams (core/params.cuh layout)."""
-    import struct
-    import pycuda.driver as cuda
-    ptr, size = module.get_global("d_params")
-    if size != struct.calcsize("=ffffI"):
-        raise RuntimeError("LBMParams host/device layout mismatch")
-    cuda.memcpy_htod(ptr, struct.pack("=ffffI", omega, fx, fy, fz, n_objects))
+def compile_kernel_v2(arch: str = "sm_70") -> str:
+    """Alias for compile_kernel() kept for test-script compatibility."""
+    return compile_kernel(arch=arch)
@@ -2,7 +2,7 @@
 """Packed ``obs`` buffer layout for force / torque / sensor telemetry.

 Usage::
-    Host code (e.g. ``ObjectManager``) and ``compiler.generate_config`` must use
+    Host code (e.g. ``ObjectManager``) and ``compiler_v2.generate_config`` must use
    the same ``dim`` and ``n_objects`` so allocation sizes match device macros.

 Design::
@@ -0,0 +1,286 @@
+# CelerisLab/lbm/block_stepper.py
+"""Opt-in periodic backend for one BlockLayout2D level, optional static circles.
+
+D2Q9/SRT/FP32 unforced double buffers. Each step refreshes every halo from
+its owner core, then pulls/collides core cells into the spare buffer, then
+swaps. Only halo pair tables are uploaded; dense ownership stays on the host.
+The caller owns the CudaContext and must close this backend before it.
+Optional ``objects`` are stationary passive circles (static linear BFL via
+receiver-local staging: halo refresh -> cut links -> pull, no copy between).
+No force/actuation readout, motion, walls, refinement, checkpoint or runtime
+parameter mutation; wall/slot/inlet fields in cfg are ignored.
+"""
+from dataclasses import replace
+from numbers import Integral
+import struct
+from threading import RLock
+
+import numpy as np
+import pycuda.driver as cuda
+
+from ..cuda import compiler_v2 as compiler
+from .blocks import NQ, BlockLayout2D, blocked_link_tables
+from .descriptors import D2Q9_EX, D2Q9_EY, FLUID, SOLID
+from .population_storage import _StoredPopulations, _encode_physical, _decode_stored
+
+MAX_GRID_DIM = 65535  # CUDA gridDim.y/z limit (rows, blocks)
+
+
+def _static_circles(layout, objects):
+    """Host tables for stationary circles; rejects before any CUDA work.
+
+    Every periodic neighbor of every fluid core is inspected, including domain
+    edges that CircleGeometry does not enumerate, so clipped or incomplete
+    geometry cannot pass. Each physical link has one receiver-owning block.
+    """
+    from ..body.geometry.circle import CircleGeometry
+
+    ny, nx = layout.shape
+    flags = np.full(ny * nx, FLUID, np.uint16)
+    for obj in objects:
+        if (type(obj.geometry) is not CircleGeometry or obj.is_sensor or obj.is_force_region
+                or obj.surface_actuation is not None
+                or any(getattr(obj.state, name) != 0 for name in ("vx", "vy", "omega"))
+                or any(vars(obj.control).values())):
+            raise ValueError("Block bodies must be stationary passive circles")
+        cx, cy = obj.geometry.center
+        radius = obj.geometry._radius
+        if not np.isfinite([cx, cy, radius]).all() or radius <= 0:
+            raise ValueError("Block circle center/radius must be finite with radius > 0")
+        mask = obj.geometry.build_flag_mask(nx, ny)
+        solid = (mask & SOLID) != 0
+        if not solid.any() or np.any(flags[solid] & SOLID):
+            raise ValueError("Block circles must be resolved and disjoint")
+        flags[solid] = mask[solid]
+    solid = ((flags & SOLID) != 0).reshape(ny, nx)
+    if solid[[0, -1]].any() or solid[:, [0, -1]].any():
+        raise ValueError("Block circles must not touch the domain edge")
+    links = [link for obj in objects
+             for link in obj.geometry.build_cut_links(nx, ny, domain_flags=flags)]
+    def source(r, q):
+        y, x = divmod(r, nx)
+        return (y + D2Q9_EY[q]) % ny, (x + D2Q9_EX[q]) % nx
+    receivers = {(int(link.fluid_idx), int(link.dir)) for link in links}
+    required = {(int(r), q) for r in np.flatnonzero(~solid.ravel())
+                for q in range(1, NQ) if solid[source(r, q)]}
+    if len(receivers) != len(links) or receivers != required:
+        raise ValueError("Cut links must cover every fluid pull from a solid exactly once")
+    tables, _ = blocked_link_tables(
+        layout, flags, np.array([link.fluid_idx for link in links], np.uint32),
+        np.array([link.dir for link in links], np.uint8),
+        np.array([link.q for link in links], np.float32),
+        np.array([link.fallback for link in links], np.uint8))
+    return dict(solid=solid, **tables)
+
+
+class PeriodicBlockStepper:
+    """G->G stepping of Q-major (9, ny, nx) postcollision populations.
+
+    On a device/launch failure the instance becomes unusable: close and
+    recreate it rather than assuming a partially advanced buffer rolled back.
+    """
+    def __init__(self, cfg, layout, *, context, objects=()):
+        if not isinstance(layout, BlockLayout2D) or layout.periodic != (True, True):
+            raise ValueError("PeriodicBlockStepper requires a doubly periodic BlockLayout2D")
+        checked = replace(cfg)
+        checked.validate()
+        if (not checked.is_d2q9 or checked.collision != "SRT"
+                or checked.streaming != "double_buffer" or checked.store_precision != "FP32"
+                or checked.les_enabled
+                or (checked.ny, checked.nx) != layout.shape):
+            raise ValueError("PeriodicBlockStepper requires matching D2Q9/SRT/FP32/"
+                             "double_buffer config without LES")
+        ly, lx = layout.local_shape
+        pairs = len(layout.halo_dst)
+        tpb = checked.threads_per_block
+        if layout.nblocks > MAX_GRID_DIM or ly > MAX_GRID_DIM or not 0 < pairs < 2**31:
+            raise ValueError("block count, block rows or halo pairs exceed launch limits")
+        objects = tuple(objects)  # an empty iterable takes the body-free path
+        self.boundary = _static_circles(layout, objects) if objects else None
+        links = 0 if self.boundary is None else len(self.boundary["cell"])
+        if links >= 2**31:
+            raise ValueError("cut links exceed launch limits")
+        self._ddf_shifting = bool(checked.ddf_shifting)
+        self.layout = layout
+        self.ctx = context
+        self._links = links
+        self._link_grid = ((links + tpb - 1) // tpb, 1)
+        # Wall/slot fields are not applied; neutralize them like the fine patch level.
+        self._local_cfg = replace(checked, nx=lx, ny=ly, y_wall_bc="bounce_back", top_wall_bc="velocity",
+                                  wall_slot_width=0, wall_normal_velocity=0.0, velocity=0.0)
+        self._block = (tpb, 1, 1)
+        self._grid = ((lx + tpb - 1) // tpb, ly, layout.nblocks)
+        self._copy_grid = ((pairs + tpb - 1) // tpb, 1)
+        self._bytes = int(np.prod(layout.blocked_shape)) * 4
+        self._lock = RLock()
+        self._status = "new"
+        self._step_count = 0
+        self._allocations = []
+        self._stream = self._module = self._copy = self._step = self._curved = None
+        self._tables = {}
+        with self.ctx.activate():
+            try:
+                free, _ = cuda.mem_get_info()
+                tables = sum(a.nbytes for k, a in (self.boundary or {}).items() if k != "solid")
+                if 2 * self._bytes + layout.halo_dst.nbytes + layout.halo_src.nbytes + tables > free:
+                    raise MemoryError("blocked buffers exceed free device memory")
+                self._stream = cuda.Stream()
+                config = self._local_cfg
+                config.validate()
+                if not np.isclose(config.omega, 1 / (3 * checked.viscosity + .5), rtol=1e-14, atol=0):
+                    raise RuntimeError("block omega/tau mismatch")
+                build = compiler.generate_config(config)
+                try:
+                    ptx = compiler.compile_kernel(self.ctx.sm_arch, source="blocks.cu", build_dir=build)
+                    module = compiler.load_module(ptx)
+                finally:
+                    compiler.cleanup_build_dir(build)
+                ptr, size = module.get_global("d_params")
+                if size != struct.calcsize("=ffffI"):
+                    raise RuntimeError("LBMParams host/device layout mismatch")
+                cuda.memcpy_htod(ptr, struct.pack("=ffffI", config.omega, 0, 0, 0, 0))
+                self._module = module
+                self._current = self._allocate(self._bytes)
+                self._spare = self._allocate(self._bytes)
+                self._dst = self._upload(layout.halo_dst)
+                self._src = self._upload(layout.halo_src)
+                self._copy = module.get_function("BlockHaloCopy")
+                self._step = module.get_function("BlockStep")
+                if self.boundary is not None:
+                    self._tables = {k: self._upload(a) for k, a in self.boundary.items() if k != "solid"}
+                    self._curved = module.get_function("BlockCurvedBoundary")
+            except BaseException:
+                try:
+                    self.close()
+                except Exception:
+                    pass  # Preserve the original construction failure.
+                raise
+
+    def _allocate(self, size):
+        pointer = cuda.mem_alloc(size)
+        self._allocations.append(pointer)
+        return pointer
+
+    def _upload(self, array):
+        array = np.ascontiguousarray(array)
+        pointer = self._allocate(array.nbytes)
+        cuda.memcpy_htod(pointer, array)
+        return pointer
+
+    def _require(self, *states):
+        if self._status not in states:
+            raise RuntimeError(f"Block stepper is {self._status}; expected {' or '.join(states)}")
+
+    @property
+    def step_count(self):
+        return self._step_count
+
+    def initialize(self, g):
+        """Upload dense G_0; the spare buffer starts NaN so stale halos cannot hide."""
+        with self._lock, self.ctx.activate():
+            self._require("new", "ready")
+            dense = np.array(g, dtype=np.float64 if self._ddf_shifting else np.float32,
+                             order="C", copy=True)
+            if dense.shape != (NQ, *self.layout.shape):
+                raise ValueError("Expected a Q-major (9, ny, nx) array")
+            fluid = np.ones(self.layout.shape, bool) if self.boundary is None else ~self.boundary["solid"]
+            if not np.isfinite(dense[:, fluid]).all() or (dense[:, fluid].sum(axis=0) <= 0).any():
+                raise ValueError("Fluid populations must be finite with positive density")
+            dense[:, ~fluid] = np.nan  # Solid cores are deliberately invalid.
+            encoded = _encode_physical(dense, self._ddf_shifting)
+            values = encoded[:, fluid]
+            if (not np.isfinite(values).all()
+                    or (values.sum(axis=0, dtype=np.float64) + int(self._ddf_shifting) <= 0).any()):
+                raise ValueError("Stored fluid populations must be finite with positive density")
+            blocked = self.layout.to_blocked(encoded)
+            spare = np.full_like(blocked, np.nan)
+            try:
+                cuda.memcpy_htod(self._current, blocked)
+                cuda.memcpy_htod(self._spare, spare)
+                self._step_count = 0
+                self._status = "ready"
+            except BaseException:
+                self._status = "failed"
+                raise
+
+    def step(self, n=1):
+        """Advance n steps: halo refresh, then core pull/collide, per step."""
+        if isinstance(n, bool) or not isinstance(n, Integral) or n < 0:
+            raise ValueError("n must be a nonnegative integer")
+        with self._lock, self.ctx.activate():
+            self._require("ready")
+            kw = dict(block=self._block, stream=self._stream)
+            try:
+                self._status = "advancing"
+                for _ in range(n):
+                    self._copy(self._current, self._dst, self._src,
+                               np.uint32(len(self.layout.halo_dst)), grid=self._copy_grid, **kw)
+                    if self._curved is not None:
+                        t = self._tables
+                        self._curved(self._current, t["block"], t["cell"], t["dir"], t["q"],
+                                     t["fallback"], np.uint32(self._links), np.uintp(0),  # no actuation
+                                     np.uintp(0), np.uintp(0), np.uintp(0),  # no observation
+                                     grid=self._link_grid, **kw)
+                    self._step(self._current, self._spare, np.uint32(self.layout.halo),
+                               self._tables.get("flags", np.uintp(0)), np.uintp(0),  # no kind
+                               grid=self._grid, **kw)
+                    self._stream.synchronize()
+                    self._current, self._spare = self._spare, self._current
+                    self._step_count += 1
+                self._status = "ready"
+            except BaseException:
+                self._status = "failed"
+                raise
+
+    def download(self):
+        """Return physical core populations (FP64 when shifted); no host cache."""
+        snapshot = self._download_raw()
+        return _decode_stored(snapshot.values[0], snapshot.shifted)
+
+    def _download_raw(self):
+        """Return one tagged dense FP32 population array without conversion."""
+        with self._lock, self.ctx.activate():
+            self._require("ready")
+            blocked = np.empty(self.layout.blocked_shape, np.float32)
+            try:
+                cuda.memcpy_dtoh(blocked, self._current)
+            except BaseException:
+                self._status = "failed"
+                raise
+            return _StoredPopulations((self.layout.to_dense(blocked),), self._ddf_shifting)
+
+    def close(self):
+        """Release owned resources; retain them for retry if context activation fails."""
+        with self._lock:
+            if self._status == "closed":
+                return
+            self._status = "failed"
+            with self.ctx.activate():
+                self._status = "closed"
+                error = None
+                try:
+                    if self._stream is not None:
+                        self._stream.synchronize()
+                except Exception as exc:
+                    error = exc
+                self._copy = self._step = self._curved = self._module = None
+                self._tables = {}
+                for pointer in self._allocations:
+                    try:
+                        pointer.free()
+                    except Exception as exc:
+                        error = error or exc
+                self._allocations.clear()
+                self._stream = None
+                if error is not None:
+                    raise error
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *exc):
+        try:
+            self.close()
+        except Exception:
+            if exc[0] is None:
+                raise
@@ -31,7 +31,7 @@ receiver-local boundary values: the solid's owner-core copy of that slot may
 be stale, which is distinct from the unique core ownership of stored state.
 No moving geometry is implied.
 """
-from dataclasses import dataclass, replace
+from dataclasses import dataclass
 from numbers import Integral
 from types import SimpleNamespace

@@ -373,84 +373,3 @@ def blocked_link_tables(layout, flags, fluid_idx, dir, q, fallback):
                  dir=columns[1].astype(np.uint8), q=columns[2].astype(np.float32),
                  fallback=columns[3].astype(np.uint8))
    return tables, mapped
-
-
-# One blocked level as FixedPatchStepper builds and launches it (any depth). Raw argument orders of the blocked.cu kernels live only here.
-
-def level_config(cfg, layout, tau):
-    """Kernel config of one blocked level: NX/NY = block window incl. halo.
-
-    Physical wall/slot/inlet fields are neutralized; a blocked root gets them
-    from its module globals (d_block_wall/d_block_channel) instead.
-    """
-    ly, lx = layout.local_shape
-    local = replace(cfg, nx=lx, ny=ly, y_wall_bc="bounce_back", top_wall_bc="velocity",
-                    wall_slot_width=0, wall_normal_velocity=0.0, velocity=0.0)
-    local.validate()
-    if not np.isclose(local.omega, 1 / tau, rtol=1e-14, atol=0):
-        raise RuntimeError("block level omega/tau mismatch")
-    return local
-
-
-def launch_halo_copy(kernel, buffer, dst, src, count, *, block, stream):
-    """BlockHaloCopy over ``count`` uploaded (dst, src) block-cell pairs."""
-    kernel(buffer, dst, src, np.uint32(count),
-           grid=((count + block[0] - 1) // block[0], 1), block=block, stream=stream)
-
-
-def launch_curved_boundary(kernel, buffer, tables, count, *, block, stream,
-                           velocity=None, rx=None, ry=None, raw=None):
-    """BlockCurvedBoundary on uploaded link tables; None = no actuation/observation."""
-    none = np.uintp(0)
-    kernel(buffer, tables["block"], tables["cell"], tables["dir"], tables["q"], tables["fallback"],
-           np.uint32(count), none if velocity is None else velocity,
-           none if rx is None else rx, none if ry is None else ry, none if raw is None else raw,
-           grid=((count + block[0] - 1) // block[0], 1), block=block, stream=stream)
-
-
-def static_circle_tables(layout, objects):
-    """Host tables for stationary circles; rejects before any CUDA work.
-
-    Every periodic neighbor of every fluid core is inspected, including domain
-    edges that CircleGeometry does not enumerate, so clipped or incomplete
-    geometry cannot pass. Each physical link has one receiver-owning block.
-    """
-    from ..body.geometry.circle import CircleGeometry
-    from .descriptors import FLUID, SOLID
-
-    ny, nx = layout.shape
-    flags = np.full(ny * nx, FLUID, np.uint16)
-    for obj in objects:
-        if (type(obj.geometry) is not CircleGeometry or obj.is_sensor or obj.is_force_region
-                or obj.surface_actuation is not None
-                or any(getattr(obj.state, name) != 0 for name in ("vx", "vy", "omega"))
-                or any(vars(obj.control).values())):
-            raise ValueError("Block bodies must be stationary passive circles")
-        cx, cy = obj.geometry.center
-        radius = obj.geometry._radius
-        if not np.isfinite([cx, cy, radius]).all() or radius <= 0:
-            raise ValueError("Block circle center/radius must be finite with radius > 0")
-        mask = obj.geometry.build_flag_mask(nx, ny)
-        solid = (mask & SOLID) != 0
-        if not solid.any() or np.any(flags[solid] & SOLID):
-            raise ValueError("Block circles must be resolved and disjoint")
-        flags[solid] = mask[solid]
-    solid = ((flags & SOLID) != 0).reshape(ny, nx)
-    if solid[[0, -1]].any() or solid[:, [0, -1]].any():
-        raise ValueError("Block circles must not touch the domain edge")
-    links = [link for obj in objects
-             for link in obj.geometry.build_cut_links(nx, ny, domain_flags=flags)]
-    def source(r, q):
-        y, x = divmod(r, nx)
-        return (y + D2Q9_EY[q]) % ny, (x + D2Q9_EX[q]) % nx
-    receivers = {(int(link.fluid_idx), int(link.dir)) for link in links}
-    required = {(int(r), q) for r in np.flatnonzero(~solid.ravel())
-                for q in range(1, NQ) if solid[source(r, q)]}
-    if len(receivers) != len(links) or receivers != required:
-        raise ValueError("Cut links must cover every fluid pull from a solid exactly once")
-    tables, _ = blocked_link_tables(
-        layout, flags, np.array([link.fluid_idx for link in links], np.uint32),
-        np.array([link.dir for link in links], np.uint8),
-        np.array([link.q for link in links], np.float32),
-        np.array([link.fallback for link in links], np.uint8))
-    return dict(solid=solid, **tables)
@@ -94,10 +94,12 @@ class CurvedLinkSoA:
        """
        self.surface_velocity = self.aperture = None
        self.jets, self.jet_shape, self.jet_id = tuple(jets), None, None
-        from ..body.jets import _surface_links
-        self.surface_velocity, self.aperture, self.jet_shape, self.jet_id = _surface_links(
+        self.surface_velocity, self.aperture = surface_link_velocity(
            objects, cfg, flags, self.fluid_idx, self.dir, self.body_id, self.rx, self.ry,
            jets=self.jets, hits=hits)
+        if self.jets:
+            from ..body.jets import jet_link_shapes
+            self.jet_shape, self.jet_id = jet_link_shapes(self.jets, hits)

    def set_jet_amplitude(self, index, amplitude, stream=None):
        """Rewrite jet rows from cached shapes; re-upload only the velocity column.
@@ -21,24 +21,7 @@ from .refinement import (ACTIVE, OUTER_GHOST, FixedPatch2D, FixedPatchSet2D,
@dataclass(frozen=True, eq=False)
 class Hierarchy:
    """configs[l]: validated LBMConfig of level l; pairs[l]: tables of (l, l+1);
-    kinds[l]: composed per-cell kind (child coverage OUTER on every non-finest level).
-
-    Index conventions (the single place they are defined):
-      * level l = 0 (root) .. depth-1; pair index p = l-1 builds level l from level p;
-      * array shapes are (ny, nx); rectangles, origins and centers are (x, y);
-      * a pair's two members are its COARSE (parent) and FINE (child) level, for every
-        pair, not absolute L0/L1 (fine_shape, fine_origin, blocked_coarse_tables, ...);
-      * refined levels (l >= 1) of FixedPatchStepper place bodies and Jets in the root
-        EDGE frame: root cell i covers [i, i+1), center i+.5. Uniform Simulation objects
-        and the static circles of a single-level (depth 1) FixedPatchStepper use the
-        integer-NODE frame (CircleGeometry.build_flag_mask at integer (x, y)). No
-        NODE<->EDGE input conversion is applied anywhere; inside the EDGE frame, refined
-        boundaries still project root-edge coordinates to fine-local cell indices with
-        the half-cell shift ((v - o)/h - .5, see prepare_fine_boundary);
-      * cell (i, j) of level l's allocated array (ghost band and holes included) covers
-        [o + h*i, o + h*(i+1)) with (o, h) = edge_frame(l); level 1 is built from the root
-        config directly, deeper levels from their parent frame (see edge_frame).
-    """
+    kinds[l]: composed per-cell kind (child coverage OUTER on every non-finest level)."""
    configs: tuple
    pairs: tuple
    kinds: tuple
@@ -47,18 +30,6 @@ class Hierarchy:
    def depth(self):
        return len(self.kinds)

-    def edge_frame(self, level):
-        """(origin (x, y), spacing) of level ``level``'s array in the root edge frame.
-
-        Each pair's fine_origin is in parent cells: zero for a set (level-wide grid),
-        the shifted window origin for a scalar FixedPatch2D.
-        """
-        origin, spacing = (0.0, 0.0), 1.0
-        for pair in self.pairs[:level]:
-            origin = tuple(o + spacing * f for o, f in zip(origin, pair.fine_origin))
-            spacing /= 2
-        return origin, spacing
-

 def is_rectangle(value):
    value = tuple(value)
@@ -115,14 +86,14 @@ def _check_chain(coarse, fine, pair):
 def leaf_partition(h):
    """Base-cell fraction published as ACTIVE over all levels (exact: powers of two).

-    Window (scalar) levels are offset by their fine_level_offset first.
+    Window (scalar) levels are offset by their fine_level_origin first.
    """
    base = np.zeros(h.kinds[0].shape)
    oy = ox = 0
    for level, kind in enumerate(h.kinds):
        if level:
            pair = h.pairs[level - 1]
-            py, px = (pair.fine_level_offset if isinstance(pair, FixedPatch2D) else (0, 0))
+            py, px = (pair.fine_level_origin if isinstance(pair, FixedPatch2D) else (0, 0))
            oy, ox = 2 * oy + py, 2 * ox + px
        y, x = np.nonzero(kind == ACTIVE)
        np.add.at(base, ((y + oy) >> level, (x + ox) >> level), 0.25 ** level)
@@ -1,47 +0,0 @@
-// CelerisLab/lbm/kernels/blocked.cu
-// One fixed-size blocked level (lbm/blocks.py layout), specialized at compile
-// time for periodic/fine, slot/reservoir or channel stepping. Separate
-// compilation entry: NX, NY are the block window including halo, so index_f
-// and linear_index address one block; blockIdx.z selects the block, whose
-// populations start at b*NQ*TOTAL_CELLS (block-contiguous, Q-major inside).
-#include <stdint.h>
-#include <math.h>
-#include "config.h"
-#include "core/precision.cuh"
-#include "core/descriptors.cuh"
-#include "core/layout.cuh"
-#include "core/params.cuh"
-#include "core/flags.cuh"
-#include "core/obs.cuh"
-#include "core/cell_kind.cuh"
-#include "operators/macro.cuh"
-#include "operators/collision_srt.cuh"
-#if COLLISION_MODEL == 2
-#include "operators/collision_mrt.cuh"
-#endif
-#include "streaming/pull_double_buffer.cuh"
-#include "boundary/curved_boundary.cuh"
-
-#if DIM != 2 || NQ != 9 || (COLLISION_MODEL != 0 && COLLISION_MODEL != 2) || STORE_PRECISION != 0 || USE_LES || STREAMING_MODEL != 0
-#error "Blocks require FP32 D2Q9 SRT or MRT double buffers without LES"
-#endif
-#ifndef BLOCK_ROOT_BC
-#error "Blocks require a generated BLOCK_ROOT_BC mode"
-#elif BLOCK_ROOT_BC < 0 || BLOCK_ROOT_BC > 2
-#error "BLOCK_ROOT_BC must be 0, 1 or 2"
-#endif
-
-#if BLOCK_ROOT_BC == 1
-#include "boundary/block_wall.cuh"
-#endif
-#include "operators/block_collision.cuh"
-
-// Kernels, in definition order: transfers, boundary staging, then the one full
-// step selected for this module (BlockStep, or BlockChannelStep for a channel root).
-#include "step/block_transfer.cu"
-#include "step/block_boundary.cu"
-#if BLOCK_ROOT_BC != 2
-#include "step/block_step.cu"
-#else
-#include "step/block_channel.cu"
-#endif
@@ -0,0 +1,206 @@
+// CelerisLab/lbm/kernels/blocks.cu
+// One periodic level of fixed-size blocks (lbm/blocks.py layout). Separate
+// compilation entry: NX, NY are the block window including halo, so index_f
+// and linear_index address one block; blockIdx.z selects the block, whose
+// populations start at b*NQ*TOTAL_CELLS (block-contiguous, Q-major inside).
+#include <stdint.h>
+#include <math.h>
+#include "config.h"
+#include "core/precision.cuh"
+#include "core/descriptors.cuh"
+#include "core/layout.cuh"
+#include "core/params.cuh"
+#include "core/flags.cuh"
+#include "core/obs.cuh"
+#include "operators/macro.cuh"
+#include "operators/collision_srt.cuh"
+#if COLLISION_MODEL == 2
+#include "operators/collision_mrt.cuh"
+#endif
+#include "streaming/pull_double_buffer.cuh"
+#include "boundary/curved_boundary.cuh"
+#include "boundary/planar_velocity.cuh"
+
+#if DIM != 2 || NQ != 9 || (COLLISION_MODEL != 0 && COLLISION_MODEL != 2) || STORE_PRECISION != 0 || USE_LES || STREAMING_MODEL != 0
+#error "Blocks require FP32 D2Q9 SRT or MRT double buffers without LES"
+#endif
+
+// Fixed-patch cell kinds, matching refinement.py ACTIVE/INNER_GHOST/OUTER_GHOST.
+constexpr unsigned char KIND_ACTIVE = 0u, KIND_INNER = 1u, KIND_OUTER = 2u;
+
+// Zero-initialized for ordinary blocks/fine levels. Set only for parent walls;
+// NX/NY remain local storage strides, never physical wall coordinates.
+struct BlockWall { unsigned int nx, ny, slot_width; float amplitude; };
+extern "C" { __constant__ BlockWall d_block_wall; }
+
+// Halo refresh: block cell ids b*TOTAL_CELLS+c, sources are always core cells.
+extern "C" __global__ void BlockHaloCopy(
+    fpxx* f, const uint64_t* dst, const uint64_t* src, unsigned int count)
+{
+    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
+    if (row >= count) return;
+    const uint64_t d = dst[row], s = src[row];
+    fpxx* to = f + (d / TOTAL_CELLS) * NQ * TOTAL_CELLS;
+    const fpxx* from = f + (s / TOTAL_CELLS) * NQ * TOTAL_CELLS;
+    for (unsigned int q = 0; q < NQ; ++q)
+        to[index_f(d % TOTAL_CELLS, q)] = from[index_f(s % TOTAL_CELLS, q)];
+}
+
+// Fixed-patch explosion into blocked fine storage: target = owner block cell
+// id of a patch-ghost cell; donors are parent block cell ids. Dense parent
+// storage is the one-window case with coarse_cells = full domain size.
+extern "C" __global__ void BlockExplode(
+    const fpxx* coarse, fpxx* fine, const uint64_t* target, const uint64_t* donor,
+    unsigned int count, unsigned long coarse_cells)
+{
+    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
+    if (row >= count) return;
+    const uint64_t t = target[row];
+    fpxx* to = fine + (t / TOTAL_CELLS) * NQ * TOTAL_CELLS;
+    const uint64_t d = donor[row];
+    const fpxx* from = coarse + (d / coarse_cells) * NQ * coarse_cells;
+    for (unsigned int q = 0; q < NQ; ++q)
+        to[index_f(t % TOTAL_CELLS, q)] = from[q * coarse_cells + d % coarse_cells];
+}
+
+// Experimental Schornbaum--Ruede (arXiv:1508.07982, Algorithm 2)
+// and Chen et al. (doi:10.1016/j.physa.2005.09.036) population interpolation.
+// Independent implementation of the mathematical stencil; no upstream source copied.
+// Stencil cells are owner cores; a disabled axis repeats its center twice.
+extern "C" __global__ void BlockExplodeProjected(
+    const fpxx* coarse, fpxx* fine, const uint64_t* target, const uint64_t* donor,
+    unsigned int count, unsigned long coarse_cells,
+    const uint64_t* neighbors, const float* offsets)
+{
+    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
+    if (row >= count) return;
+    const uint64_t t = target[row], d = donor[row];
+    fpxx* to = fine + (t / TOTAL_CELLS) * NQ * TOTAL_CELLS;
+    for (unsigned int q = 0; q < NQ; ++q) {
+        float v[4];
+        for (unsigned int a = 0; a < 4; ++a) {
+            const uint64_t n = neighbors[4ul * row + a];
+            v[a] = coarse[(n / coarse_cells) * NQ * coarse_cells
+                          + q * coarse_cells + n % coarse_cells];
+        }
+        float gx = .5f * (v[1] - v[0]), gy = .5f * (v[3] - v[2]);
+        const float ex = d_cx[q], ey = d_cy[q];
+        if (q != 0) {
+            const float along = (ex * gx + ey * gy) / (ex * ex + ey * ey);
+            gx -= ex * along;
+            gy -= ey * along;
+        }
+        to[index_f(t % TOTAL_CELLS, q)] =
+            coarse[(d / coarse_cells) * NQ * coarse_cells + q * coarse_cells + d % coarse_cells]
+            + offsets[2ul * row] * gx + offsets[2ul * row + 1] * gy;
+    }
+}
+
+// Targets are receiver-local SOURCE slots for the immediately following pull.
+// Keep child summation order identical to PatchCoarseStep; no collision here.
+extern "C" __global__ void BlockCoalesce(
+    const fpxx* fine, fpxx* coarse, const uint64_t* target, const uint64_t* children,
+    unsigned int count)
+{
+    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
+    if (row >= count) return;
+    const uint64_t* s = children + 4ul * row;
+    coarse[target[row]] = 0.25f * (fine[s[0]] + fine[s[1]] + fine[s[2]] + fine[s[3]]);
+}
+
+// Static cut links, receiver-local staging: cell is the receiver's local core
+// cell, so the existing BFL writes the solid-source window image that this
+// block's next pull reads (no halo copy between this kernel and BlockStep).
+// surface_velocity (nullable, 2 floats per link): prescribed fluid velocity on
+// fixed geometry, same convention as CurvedBoundaryKernel. Optional raw rows
+// (jx,jy,torque,dm) are unique physical links, never body-observation slots.
+extern "C" __global__ void BlockCurvedBoundary(
+    fpxx* f, const uint32_t* block, const uint32_t* cell, const unsigned char* dir,
+    const float* q, const unsigned char* fallback, unsigned int count,
+    const float* surface_velocity, const float* rx, const float* ry, float* raw)
+{
+    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
+    if (row >= count) return;
+    float Uw = 0.0f, Vw = 0.0f;
+    if (surface_velocity != nullptr) {
+        const float u = surface_velocity[2 * row], v = surface_velocity[2 * row + 1];
+        if (u != 0.0f || v != 0.0f) { Uw = u; Vw = v; }
+    }
+    apply_bouzidi_link(dir[row], q[row], cell[row], f + (unsigned long)block[row] * NQ * TOTAL_CELLS,
+                       Uw, Vw, rx ? rx[row] : 0.0f, ry ? ry[row] : 0.0f,
+                       fallback[row], nullptr, -1, raw ? raw + 4ul * row : nullptr,
+                       raw ? raw + 4ul * row + 3 : nullptr);
+}
+
+// Same-stream BFL -> this row-owned accumulation -> pull step. No atomics,
+// no cross-row reduction and no history allocation proportional to substeps.
+extern "C" __global__ void BlockAccumulateExchange(
+    const float* raw, double* integral, unsigned int count, double h, double dt)
+{
+    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
+    if (row >= count) return;
+    const unsigned long k = 4ul * row;
+    const double impulse = h * dt;
+    integral[k] += impulse * (double)raw[k];
+    integral[k + 1] += impulse * (double)raw[k + 1];
+    integral[k + 2] += h * impulse * (double)raw[k + 2];
+    integral[k + 3] += h * h * (double)raw[k + 3];
+}
+
+// Pull/collide/store core cells only; halo cells of out are left untouched.
+// flags (block cell ids, nullable): solid cores store NaN and never collide.
+// kind (block cell ids, nullable = all ACTIVE): OUTER skips the pull and
+// stores NaN, INNER pulls without collision, ACTIVE pulls and collides.
+extern "C" __global__ void BlockStep(const fpxx* in, fpxx* out, unsigned int halo,
+                                     const uint16_t* flags, const unsigned char* kind) {
+    unsigned int x, y; unsigned long k;
+    index_from_thread(x, y, k);
+    if (x < halo || y < halo || x >= NX - halo || y >= NY - halo) return;
+    const unsigned long base = (unsigned long)blockIdx.z * NQ * TOTAL_CELLS;
+    const unsigned long cell = (unsigned long)blockIdx.z * TOTAL_CELLS + k;
+    const unsigned char cell_kind = kind != nullptr ? kind[cell] : KIND_ACTIVE;
+    float f[NQ];
+    if (cell_kind == KIND_OUTER || (flags != nullptr && is_solid(flags[cell]))) {
+        for (int q = 0; q < NQ; ++q) f[q] = NAN;
+    } else {
+        unsigned long j[NQ];
+        compute_neighbors(x, y, j);  // Core neighbors stay inside the window: no wrap.
+        stream_pull_load(k, f, in + base, j);
+        if (cell_kind == KIND_ACTIVE) {
+            if (d_block_wall.nx != 0) {
+                const unsigned int bx = NX - 2 * halo, by = NY - 2 * halo;
+                const unsigned int nbx = d_block_wall.nx / bx;
+                const unsigned int gx = (blockIdx.z % nbx) * bx + x - halo;
+                const unsigned int gy = (blockIdx.z / nbx) * by + y - halo;
+                if (gy == 1)
+                    reconstruct_bottom_velocity(f, 0.0f, bottom_wall_normal_velocity(
+                        gx, d_block_wall.nx, d_block_wall.slot_width, d_block_wall.amplitude));
+                else if (gy == d_block_wall.ny - 2) reconstruct_top_reservoir(f, RHO);
+            }
+#if COLLISION_MODEL == 2
+            // Uniform MRT sequence (collide_dispatch, zero forcing); SRT path below unchanged.
+            float rho, ux, uy;
+            compute_rho_u(f, rho, ux, uy);
+            collide_mrt_no_force(f, rho, ux, uy,
+                fminf(OMEGA_COLLISION_MAX, fmaxf(OMEGA_COLLISION_MIN, d_params.omega)));
+#else
+            float rho, ux, uy, feq[NQ];
+            compute_rho_u(f, rho, ux, uy);
+#if USE_DDF_SHIFTING
+            // Preserve the density perturbation before adding the unit reference.
+            const float drho = f[0]+f[1]+f[2]+f[3]+f[4]+f[5]+f[6]+f[7]+f[8];
+            const float jx = f[1]-f[2]+f[5]-f[6]+f[7]-f[8];
+            const float jy = f[3]-f[4]+f[5]-f[6]-f[7]+f[8];
+            compute_feq_from_momentum(drho, jx, jy, ux, uy, feq);
+#else
+            compute_feq(rho, ux, uy, feq);
+#endif
+            collide_srt_no_force(f, feq, d_params.omega);
+#endif
+        }
+    }
+    stream_pull_store(k, f, out + base);
+}
+
+// Root channel closure; fine and periodic levels keep BlockStep.
+#include "step/block_channel.cu"
@@ -1,20 +0,0 @@
-// Physical parent walls; NX/NY remain local storage strides.
-#ifndef CELERIS_BOUNDARY_BLOCK_WALL_CUH
-#define CELERIS_BOUNDARY_BLOCK_WALL_CUH
-
-#include "planar_velocity.cuh"
-
-struct BlockWall { unsigned int nx, ny, slot_width; float amplitude; };
-extern "C" { __constant__ BlockWall d_block_wall; }
-
-__device__ __forceinline__ void block_wall(float* f, unsigned int x, unsigned int y,
-                                         unsigned int halo) {
-    unsigned int gx, gy;
-    block_global_xy(x, y, halo, d_block_wall.nx, gx, gy);
-    if (gy == 1)
-        reconstruct_bottom_velocity(f, 0.0f, bottom_wall_normal_velocity(
-            gx, d_block_wall.nx, d_block_wall.slot_width, d_block_wall.amplitude));
-    else if (gy == d_block_wall.ny - 2) reconstruct_top_reservoir(f, RHO);
-}
-
-#endif
@@ -15,13 +15,10 @@
 // D2Q9 y-wall half-way bounce-back for pull double-buffer streaming.
 // ---------------------------------------------------------------------------
 #if LATTICE_MODEL == LATTICE_D2Q9
-// Rows 1 and top_row (= physical ny - 2) next to walls at rows 0 and ny - 1.
-// y is the physical row; k addresses the node in fi_in (a block window may
-// have NX/NY strides different from the physical extent).
-__device__ inline void apply_wall_bb_d2q9_rows(unsigned int y, unsigned int top_row,
-                                               float* __restrict__ f,
-                                               const fpxx* __restrict__ fi_in,
-                                               unsigned long k)
+__device__ inline void apply_wall_bb_d2q9_y_pull(unsigned int y,
+                                                  float* __restrict__ f,
+                                                  const fpxx* __restrict__ fi_in,
+                                                  unsigned long k)
 {
    if (y == 1) {
        // Directions sourced from y=0 wall in pull step: +y, +x+y, -x+y.
@@ -29,22 +26,14 @@ __device__ inline void apply_wall_bb_d2q9_rows(unsigned int y, unsigned int top_
        f[5] = load_ddf(fi_in, index_f(k, 6u));
        f[8] = load_ddf(fi_in, index_f(k, 7u));
    }
-    else if (y == top_row) {
-        // Directions sourced from the y = ny - 1 wall in pull step: -y, -x-y, +x-y.
+    else if (y == (unsigned int)(NY - 2)) {
+        // Directions sourced from y=NY-1 wall in pull step: -y, -x-y, +x-y.
        f[4] = load_ddf(fi_in, index_f(k, 3u));
        f[6] = load_ddf(fi_in, index_f(k, 5u));
        f[7] = load_ddf(fi_in, index_f(k, 8u));
    }
 }

-__device__ inline void apply_wall_bb_d2q9_y_pull(unsigned int y,
-                                                  float* __restrict__ f,
-                                                  const fpxx* __restrict__ fi_in,
-                                                  unsigned long k)
-{
-    apply_wall_bb_d2q9_rows(y, (unsigned int)(NY - 2), f, fi_in, k);
-}
-
 // ---------------------------------------------------------------------------
 // D2Q9 y-wall free-slip (specular reflection) for pull double-buffer streaming.
 //
@@ -8,9 +8,6 @@
 #define STORE_PRECISION  0
 #define USE_DDF_SHIFTING 0

-// Internal blocked root mode: 0=periodic/fine, 1=slot/reservoir, 2=channel.
-#define BLOCK_ROOT_BC 0
-
 #define USE_LES 0
 #define LES_CS  0.160000f
 #define LES_CLOSED_FORM 1
@@ -1,16 +0,0 @@
-// CelerisLab – core/cell_kind.cuh
-// Fixed-patch cell kinds stored per block cell, matching lbm/refinement.py
-// ACTIVE/INNER_GHOST/OUTER_GHOST: OUTER is invalid (stores NaN, no pull),
-// INNER streams without collision, ACTIVE pulls and collides.
-#ifndef CELERIS_CORE_CELL_KIND_CUH
-#define CELERIS_CORE_CELL_KIND_CUH
-
-constexpr unsigned char KIND_ACTIVE = 0u, KIND_INNER = 1u, KIND_OUTER = 2u;
-
-// NaN populations mark a cell without valid state (OUTER ghost or solid core).
-__device__ __forceinline__ void fill_invalid(float* f)
-{
-    for (int q = 0; q < NQ; ++q) f[q] = NAN;
-}
-
-#endif
@@ -53,18 +53,6 @@ __device__ __forceinline__ void index_from_thread(unsigned int& x,
    k = (unsigned long)y * NX + x;
 }

-// Global cell of window cell (x, y) in block blockIdx.z of a row-major,
-// full-domain block layout domain_nx cells wide. NX/NY are the local strides.
-__device__ __forceinline__ void block_global_xy(unsigned int x, unsigned int y,
-                                                unsigned int halo, unsigned int domain_nx,
-                                                unsigned int& gx, unsigned int& gy)
-{
-    const unsigned int bx = NX - 2 * halo, by = NY - 2 * halo;
-    const unsigned int nbx = domain_nx / bx;
-    gx = (blockIdx.z % nbx) * bx + x - halo;
-    gy = (blockIdx.z / nbx) * by + y - halo;
-}
-
 // ---------------------------------------------------------------------------
 // Neighbor list (periodic BCs via modular arithmetic)
 // ---------------------------------------------------------------------------
@@ -44,4 +44,16 @@ struct RigidBodyControl2D {
    int   control_mode;           // 0=force, 1=velocity, 2=displacement
 };

+// ============================================================================
+// Halo plan (multi-GPU / AMR 预留)
+// ============================================================================
+struct HaloPlan {
+    int       face_id;
+    int       peer_gpu;
+    int       level;
+    size_t    count;
+    int*      send_idx;
+    int*      recv_idx;
+};
+
 #endif // CELERIS_CORE_PARAMS_CUH
@@ -1,4 +1,4 @@
-// CelerisLab – uniform.cu
+// CelerisLab – kernel_v2.cu
 // ============================================================================
 // Compilation entry point (Stage 2 Architecture).
 //
@@ -7,7 +7,7 @@
 //
 // File layout (each <300 lines):
 //   operators/helpers.cuh  — bounce_back_swap, collide_dispatch
-//   step/init_flow.cu      — InitTubeFlow, InitEsoPull
+//   step/init_flow.cu      — InitTubeFlow_v2, InitEsoPull
 //   step/one_step_double.cu — OneStep (pull double-buffer)
 //   step/one_step_esopull.cu — EsoPullStep (single-buffer)
 //   step/aux_kernels.cu    — CurvedBoundaryKernel, SensorKernel
@@ -1,43 +0,0 @@
-// The one collision extension point of the blocked steps (BlockStep, BlockChannelStep).
-// Compile-time selection: MRT shares one path; SRT keeps two explicit arithmetic
-// variants, unchanged from their original kernels:
-//   level  (BLOCK_ROOT_BC 0/1): unclamped omega, no forcing term; shifted storage
-//          builds feq from the retained drho and direct momentum (numerical variant);
-//   channel (BLOCK_ROOT_BC 2): matches uniform OneStep SRT arithmetic (omega clamp,
-//          zero Guo forcing term, compute_feq).
-#ifndef CELERIS_OPERATORS_BLOCK_COLLISION_CUH
-#define CELERIS_OPERATORS_BLOCK_COLLISION_CUH
-
-#include "../core/params.cuh"
-#include "macro.cuh"
-#include "collision_srt.cuh"
-#if COLLISION_MODEL == 2
-#include "collision_mrt.cuh"
-#endif
-
-__device__ __forceinline__ void block_collide(float* f) {
-#if COLLISION_MODEL == 2
-    collide_mrt_unforced(f);
-#elif BLOCK_ROOT_BC == 2
-    float rho, ux, uy, feq[NQ], forcing[NQ] = {};
-    compute_rho_u(f, rho, ux, uy);
-    compute_feq(rho, ux, uy, feq);
-    const float omega = fminf(OMEGA_COLLISION_MAX, fmaxf(OMEGA_COLLISION_MIN, d_params.omega));
-    collide_srt(f, feq, forcing, omega);
-#else
-    float rho, ux, uy, feq[NQ];
-    compute_rho_u(f, rho, ux, uy);
-#if USE_DDF_SHIFTING
-    // Preserve the density perturbation before adding the unit reference.
-    const float drho = f[0]+f[1]+f[2]+f[3]+f[4]+f[5]+f[6]+f[7]+f[8];
-    const float jx = f[1]-f[2]+f[5]-f[6]+f[7]-f[8];
-    const float jy = f[3]-f[4]+f[5]-f[6]-f[7]+f[8];
-    compute_feq_from_momentum(drho, jx, jy, ux, uy, feq);
-#else
-    compute_feq(rho, ux, uy, feq);
-#endif
-    collide_srt_no_force(f, feq, d_params.omega);
-#endif
-}
-
-#endif
@@ -1,14 +0,0 @@
-// CelerisLab – operators/bounce_back_swap.cuh
-// Full-way bounce-back at a wall node: swap every opposite pair (paired
-// order q, q+1). Needs only NQ; shared by the uniform and blocked steps.
-#ifndef CELERIS_OPERATORS_BOUNCE_BACK_SWAP_CUH
-#define CELERIS_OPERATORS_BOUNCE_BACK_SWAP_CUH
-
-__device__ __forceinline__ void bounce_back_swap(float* f) {
-    #pragma unroll
-    for (int i = 1; i < NQ; i += 2) {
-        float t = f[i]; f[i] = f[i+1]; f[i+1] = t;
-    }
-}
-
-#endif
@@ -114,16 +114,6 @@ __device__ __forceinline__ void collide_mrt_no_force(float* __restrict__ g,
    collide_mrt(g, rho, ux, uy, Fin, omega);
 }

-// Unforced MRT with the clamped device omega, as in collide_dispatch with zero
-// forcing. Shared by the blocked, channel-root and dense-patch collisions.
-__device__ __forceinline__ void collide_mrt_unforced(float* __restrict__ g)
-{
-    float rho, ux, uy;
-    compute_rho_u(g, rho, ux, uy);
-    collide_mrt_no_force(g, rho, ux, uy,
-        fminf(OMEGA_COLLISION_MAX, fmaxf(OMEGA_COLLISION_MIN, d_params.omega)));
-}
-
 #elif LATTICE_MODEL == LATTICE_D3Q19

 __device__ __forceinline__ void collide_mrt(float* __restrict__ g,
@@ -57,4 +57,25 @@ __device__ __forceinline__ void collide_trt(float* __restrict__ f,
    }
 }

+__device__ __forceinline__ void collide_trt_no_force(float* __restrict__ f,
+                                                     const float* __restrict__ feq,
+                                                     float omega)
+{
+    const float wp = omega;
+    const float wm = compute_omega_minus(wp);
+
+    f[0] = f[0] - wp * (f[0] - feq[0]);
+
+    #pragma unroll
+    for (int i = 1; i < NQ; i += 2) {
+        const int ib = i + 1;
+        const float fi = f[i], fb = f[ib];
+        const float feqi = feq[i], feqb = feq[ib];
+        const float delta_s = 0.5f * ((fi - feqi) + (fb - feqb));
+        const float delta_a = 0.5f * ((fi - feqi) - (fb - feqb));
+        f[i]  = fi  - wp * delta_s - wm * delta_a;
+        f[ib] = fb  - wp * delta_s + wm * delta_a;
+    }
+}
+
 #endif // CELERIS_OPERATORS_COLLISION_TRT_CUH
@@ -5,7 +5,12 @@
 #ifndef CELERIS_OPERATORS_HELPERS_CUH
 #define CELERIS_OPERATORS_HELPERS_CUH

-#include "bounce_back_swap.cuh"
+__device__ __forceinline__ void bounce_back_swap(float* f) {
+    #pragma unroll
+    for (int i = 1; i < NQ; i += 2) {
+        float t = f[i]; f[i] = f[i+1]; f[i+1] = t;
+    }
+}

 __device__ __forceinline__ void collide_dispatch(
    float* f, float rho_n,
@@ -0,0 +1,130 @@
+// Fixed, unforced internal 2:1 patch. Separate compilation entry;
+// default OneStep/channel kernels are deliberately not included.
+#include <stdint.h>
+#include <math.h>
+#include "config.h"
+#include "core/precision.cuh"
+#include "core/descriptors.cuh"
+#include "core/layout.cuh"
+#include "core/params.cuh"
+#include "core/flags.cuh"
+#include "core/obs.cuh"
+#include "operators/macro.cuh"
+#include "operators/collision_srt.cuh"
+#if COLLISION_MODEL == 2
+#include "operators/collision_mrt.cuh"
+#endif
+#include "streaming/pull_double_buffer.cuh"
+#include "boundary/planar_velocity.cuh"
+#include "boundary/curved_boundary.cuh"
+
+extern "C" {
+#include "step/curved_double.cu"
+}
+
+#if DIM != 2 || NQ != 9 || (COLLISION_MODEL != 0 && COLLISION_MODEL != 2) || STORE_PRECISION != 0 || USE_DDF_SHIFTING || USE_LES || STREAMING_MODEL != 0
+#error "Refinement requires unshifted FP32 D2Q9 SRT or MRT double buffers without LES"
+#endif
+
+// Enabled only in the validation build; no diagnostic arrays in normal builds.
+#ifdef CELERIS_REFINEMENT_AUDIT
+extern "C" {
+__device__ unsigned int collision_counts[TOTAL_CELLS];
+__device__ unsigned int explosion_counts[TOTAL_CELLS];
+__device__ unsigned int incoming_counts[NQ * TOTAL_CELLS];
+}
+#endif
+
+__device__ __forceinline__ void patch_collide(float* f, unsigned long k) {
+#if COLLISION_MODEL == 2
+    float rho, ux, uy;
+    compute_rho_u(f, rho, ux, uy);
+    collide_mrt_no_force(f, rho, ux, uy,
+        fminf(OMEGA_COLLISION_MAX, fmaxf(OMEGA_COLLISION_MIN, d_params.omega)));
+#else
+    float rho, ux, uy, feq[NQ];
+    compute_rho_u(f, rho, ux, uy);
+    compute_feq(rho, ux, uy, feq);
+    collide_srt_no_force(f, feq, d_params.omega);
+#endif
+#ifdef CELERIS_REFINEMENT_AUDIT
+    ++collision_counts[k];
+#endif
+}
+
+extern "C" __global__ void PatchExplode(
+    const float* coarse, float* fine, const uint32_t* targets,
+    const uint32_t* donors, unsigned int count, unsigned long coarse_cells)
+{
+    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
+    if (row >= count) return;
+    for (unsigned int q = 0; q < NQ; ++q)
+        fine[index_f(targets[row], q)] = coarse[q * coarse_cells + donors[row]];
+#ifdef CELERIS_REFINEMENT_AUDIT
+    ++explosion_counts[targets[row]];
+#endif
+}
+
+extern "C" __global__ void PatchFineStep(const float* in, float* out, const uint16_t* flags) {
+    unsigned int x, y; unsigned long k;
+    index_from_thread(x, y, k);
+    if (x >= NX || y >= NY) return;
+    float f[NQ];
+    // Outer halo is invalid after EACH hop, including when reusing the buffer
+    // that held explosion data. Inner halo only streams; it never collides.
+    if (x < 2 || y < 2 || x >= NX - 2 || y >= NY - 2
+            || (flags != nullptr && is_solid(flags[k]))) {
+        for (int q = 0; q < NQ; ++q) f[q] = NAN;
+    } else {
+        unsigned long j[NQ];
+        compute_neighbors(x, y, j);  // These interior targets never wrap.
+        stream_pull_load(k, f, in, j);
+        if (x >= 4 && y >= 4 && x < NX - 4 && y < NY - 4)
+            patch_collide(f, k);
+    }
+    stream_pull_store(k, f, out);
+}
+
+extern "C" __global__ void PatchCoarseStep(
+    const float* in, float* out, const float* fine,
+    const uint64_t* targets, const uint64_t* children, unsigned int count,
+    unsigned int x0, unsigned int y0, unsigned int x1, unsigned int y1,
+    unsigned int wall_aware)
+{
+    unsigned int x, y; unsigned long k;
+    index_from_thread(x, y, k);
+    if (x >= NX || y >= NY) return;
+    // Both buffers were poisoned at initialization; covered cells stay invalid.
+    if (x >= x0 && x < x1 && y >= y0 && y < y1) return;
+    if (wall_aware && (y == 0 || y == NY - 1)) return;
+    unsigned long j[NQ];
+    compute_neighbors(x, y, j);
+    float f[NQ];
+    stream_pull_load(k, f, in, j);
+    if (x >= x0 - 1 && x <= x1 && y >= y0 - 1 && y <= y1) {
+        for (unsigned int q = 1; q < NQ; ++q) {
+            uint64_t slot = index_f(k, q);
+            unsigned int lo = 0, hi = count;
+            while (lo < hi) {
+                unsigned int mid = lo + (hi - lo) / 2;
+                if (targets[mid] < slot) lo = mid + 1;
+                else hi = mid;
+            }
+            if (lo < count && targets[lo] == slot) {
+                const uint64_t* s = children + 4ul * lo;
+                f[q] = 0.25f * (fine[s[0]] + fine[s[1]] + fine[s[2]] + fine[s[3]]);
+#ifdef CELERIS_REFINEMENT_AUDIT
+                ++incoming_counts[slot];
+#endif
+            }
+        }
+    }
+    // Includes the eight grazing-corner links with UNCOVERED coarse donors.
+    // All replacement arrivals must be in place before computing rho/u.
+    if (wall_aware) {
+        if (y == 1) reconstruct_bottom_velocity(f, 0.0f, bottom_wall_normal_velocity(x));
+        else if (y == NY - 2) reconstruct_top_reservoir(f, RHO);
+    }
+    patch_collide(f, k);
+    stream_pull_store(k, f, out);
+}
@@ -3,7 +3,7 @@
 // ApplyEsoPullCurvedKernel, SensorKernel, ForceRegionKernel.
 //
 // Launched on compact lists (n_curved / n_sensor / n_cells threads).
-// Included by uniform.cu.  No standalone compilation.
+// Included by kernel_v2.cu.  No standalone compilation.
 //
 // Double-buffer mode:
 //   CurvedBoundaryKernel runs BEFORE OneStep.
@@ -1,42 +0,0 @@
-// CelerisLab/lbm/kernels/step/block_boundary.cu
-// Receiver-local curved-boundary staging and its raw-exchange accumulation.
-// Part of the blocked.cu compilation entry; relies on its includes and guards.
-
-// Static cut links, receiver-local staging: cell is the receiver's local core
-// cell, so the existing BFL writes the solid-source window image that this
-// block's next pull reads (no halo copy between this kernel and BlockStep).
-// surface_velocity (nullable, 2 floats per link): prescribed fluid velocity on
-// fixed geometry, same convention as CurvedBoundaryKernel. Optional raw rows
-// (jx,jy,torque,dm) are unique physical links, never body-observation slots.
-extern "C" __global__ void BlockCurvedBoundary(
-    fpxx* f, const uint32_t* block, const uint32_t* cell, const unsigned char* dir,
-    const float* q, const unsigned char* fallback, unsigned int count,
-    const float* surface_velocity, const float* rx, const float* ry, float* raw)
-{
-    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
-    if (row >= count) return;
-    float Uw = 0.0f, Vw = 0.0f;
-    if (surface_velocity != nullptr) {
-        const float u = surface_velocity[2 * row], v = surface_velocity[2 * row + 1];
-        if (u != 0.0f || v != 0.0f) { Uw = u; Vw = v; }
-    }
-    apply_bouzidi_link(dir[row], q[row], cell[row], f + (unsigned long)block[row] * NQ * TOTAL_CELLS,
-                       Uw, Vw, rx ? rx[row] : 0.0f, ry ? ry[row] : 0.0f,
-                       fallback[row], nullptr, -1, raw ? raw + 4ul * row : nullptr,
-                       raw ? raw + 4ul * row + 3 : nullptr);
-}
-
-// Same-stream BFL -> this row-owned accumulation -> pull step. No atomics,
-// no cross-row reduction and no history allocation proportional to substeps.
-extern "C" __global__ void BlockAccumulateExchange(
-    const float* raw, double* integral, unsigned int count, double h, double dt)
-{
-    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
-    if (row >= count) return;
-    const unsigned long k = 4ul * row;
-    const double impulse = h * dt;
-    integral[k] += impulse * (double)raw[k];
-    integral[k + 1] += impulse * (double)raw[k + 1];
-    integral[k + 2] += h * impulse * (double)raw[k + 2];
-    integral[k + 3] += h * h * (double)raw[k + 3];
-}
@@ -1,12 +1,5 @@
-#if BLOCK_ROOT_BC != 2
-#error "BlockChannelStep requires BLOCK_ROOT_BC == 2"
-#endif
-
 // Root-only blocked counterpart of OneStep's D2Q9/SRT channel closure.
 // NX/NY are LOCAL strides; all physical boundary decisions use global coordinates.
-#include "../core/cell_kind.cuh"
-#include "../operators/bounce_back_swap.cuh"
-#include "../boundary/bounce_back.cuh"
 #include "../boundary/inlet/common.cuh"
 #include "../boundary/inlet/zou_he_local.cuh"
 #include "../boundary/outlet/pressure_neq.cuh"
@@ -22,13 +15,15 @@ extern "C" __global__ void BlockChannelStep(
    unsigned int x, y; unsigned long k;
    index_from_thread(x, y, k);
    if (x < halo || y < halo || x >= NX - halo || y >= NY - halo) return;
-    unsigned int gx, gy;
-    block_global_xy(x, y, halo, d_block_channel.nx, gx, gy);
+    const unsigned int bx = NX - 2 * halo, by = NY - 2 * halo;
+    const unsigned int nbx = d_block_channel.nx / bx;
+    const unsigned int gx = (blockIdx.z % nbx) * bx + x - halo;
+    const unsigned int gy = (blockIdx.z / nbx) * by + y - halo;
    const unsigned long base = (unsigned long)blockIdx.z * NQ * TOTAL_CELLS;
    const unsigned long cell = (unsigned long)blockIdx.z * TOTAL_CELLS + k;
    float f[NQ];
    if (kind != nullptr && kind[cell] == KIND_OUTER) {
-        fill_invalid(f);
+        for (int q = 0; q < NQ; ++q) f[q] = NAN;
    } else {
        unsigned long j[NQ];
        compute_neighbors(x, y, j);
@@ -36,19 +31,44 @@ extern "C" __global__ void BlockChannelStep(
        const bool wall = gy == 0 || gy == d_block_channel.ny - 1;
        const bool outlet = gx == d_block_channel.nx - 1;
        if (wall) {
-            bounce_back_swap(f);  // Wall owns all four corners; ghosts never collide.
+            // Wall owns all four corners. Ghosts pull and swap, never collide.
+            for (int q = 1; q < NQ; q += 2) {
+                const float t = f[q]; f[q] = f[q + 1]; f[q + 1] = t;
+            }
        } else if (gx == 0) {
            apply_zou_he_left_velocity_inlet_d2q9(
                f, inlet_target_u((float)gy, d_block_channel.ny, d_block_channel.velocity), 0.0f);
        } else if (outlet) {
            // Immutable input donor, receiver-local (also valid across a block seam).
            float donor[NQ];
-            load_node(in + base, k - 1, donor);
+            for (int q = 0; q < NQ; ++q)
+                donor[q] = load_ddf(in + base, index_f(k - 1, q));
            apply_pressure_outlet_d2q9(f, donor);
-        } else {
-            apply_wall_bb_d2q9_rows(gy, d_block_channel.ny - 2, f, in + base, k);
+        } else if (gy == 1) {
+            f[3] = load_ddf(in + base, index_f(k, 4));
+            f[5] = load_ddf(in + base, index_f(k, 6));
+            f[8] = load_ddf(in + base, index_f(k, 7));
+        } else if (gy == d_block_channel.ny - 2) {
+            f[4] = load_ddf(in + base, index_f(k, 3));
+            f[6] = load_ddf(in + base, index_f(k, 5));
+            f[7] = load_ddf(in + base, index_f(k, 8));
+        }
+        if (!wall && !outlet) {
+#if COLLISION_MODEL == 2
+            // Match uniform MRT arithmetic (collide_dispatch, zero forcing).
+            float rho, ux, uy;
+            compute_rho_u(f, rho, ux, uy);
+            collide_mrt_no_force(f, rho, ux, uy,
+                fminf(OMEGA_COLLISION_MAX, fmaxf(OMEGA_COLLISION_MIN, d_params.omega)));
+#else
+            // Match uniform SRT arithmetic, including its shifted equilibrium path.
+            float rho, ux, uy, feq[NQ], forcing[NQ] = {};
+            compute_rho_u(f, rho, ux, uy);
+            compute_feq(rho, ux, uy, feq);
+            const float omega = fminf(OMEGA_COLLISION_MAX, fmaxf(OMEGA_COLLISION_MIN, d_params.omega));
+            collide_srt(f, feq, forcing, omega);
+#endif
        }
-        if (!wall && !outlet) block_collide(f);
    }
    stream_pull_store(k, f, out + base);
 }
@@ -1,35 +0,0 @@
-// CelerisLab/lbm/kernels/step/block_step.cu
-// Generic blocked pull/collide/store for periodic/fine and slot/reservoir levels.
-// Part of the blocked.cu compilation entry; relies on its includes and guards.
-#if BLOCK_ROOT_BC == 2
-#error "BlockStep is not built for the channel root (BLOCK_ROOT_BC == 2)"
-#endif
-
-// Pull/collide/store core cells only; halo cells of out are left untouched.
-// flags (block cell ids, nullable): solid cores store NaN and never collide.
-// kind (block cell ids, nullable = all ACTIVE): OUTER skips the pull and
-// stores NaN, INNER pulls without collision, ACTIVE pulls and collides.
-extern "C" __global__ void BlockStep(const fpxx* in, fpxx* out, unsigned int halo,
-                                     const uint16_t* flags, const unsigned char* kind) {
-    unsigned int x, y; unsigned long k;
-    index_from_thread(x, y, k);
-    if (x < halo || y < halo || x >= NX - halo || y >= NY - halo) return;
-    const unsigned long base = (unsigned long)blockIdx.z * NQ * TOTAL_CELLS;
-    const unsigned long cell = (unsigned long)blockIdx.z * TOTAL_CELLS + k;
-    const unsigned char cell_kind = kind != nullptr ? kind[cell] : KIND_ACTIVE;
-    float f[NQ];
-    if (cell_kind == KIND_OUTER || (flags != nullptr && is_solid(flags[cell]))) {
-        fill_invalid(f);
-    } else {
-        unsigned long j[NQ];
-        compute_neighbors(x, y, j);  // Core neighbors stay inside the window: no wrap.
-        stream_pull_load(k, f, in + base, j);
-        if (cell_kind == KIND_ACTIVE) {
-#if BLOCK_ROOT_BC == 1
-            block_wall(f, x, y, halo);
-#endif
-            block_collide(f);
-        }
-    }
-    stream_pull_store(k, f, out + base);
-}
@@ -1,78 +0,0 @@
-// CelerisLab/lbm/kernels/step/block_transfer.cu
-// Same-level halo refresh and 2:1 explosion/coalescence between blocked levels.
-// Part of the blocked.cu compilation entry; relies on its includes and guards.
-
-// Halo refresh: block cell ids b*TOTAL_CELLS+c, sources are always core cells.
-extern "C" __global__ void BlockHaloCopy(
-    fpxx* f, const uint64_t* dst, const uint64_t* src, unsigned int count)
-{
-    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
-    if (row >= count) return;
-    const uint64_t d = dst[row], s = src[row];
-    fpxx* to = f + (d / TOTAL_CELLS) * NQ * TOTAL_CELLS;
-    const fpxx* from = f + (s / TOTAL_CELLS) * NQ * TOTAL_CELLS;
-    for (unsigned int q = 0; q < NQ; ++q)
-        to[index_f(d % TOTAL_CELLS, q)] = from[index_f(s % TOTAL_CELLS, q)];
-}
-
-// Fixed-patch explosion into blocked fine storage: target = owner block cell
-// id of a patch-ghost cell; donors are parent block cell ids and coarse_cells
-// is the parent's block window size.
-extern "C" __global__ void BlockExplode(
-    const fpxx* coarse, fpxx* fine, const uint64_t* target, const uint64_t* donor,
-    unsigned int count, unsigned long coarse_cells)
-{
-    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
-    if (row >= count) return;
-    const uint64_t t = target[row];
-    fpxx* to = fine + (t / TOTAL_CELLS) * NQ * TOTAL_CELLS;
-    const uint64_t d = donor[row];
-    const fpxx* from = coarse + (d / coarse_cells) * NQ * coarse_cells;
-    for (unsigned int q = 0; q < NQ; ++q)
-        to[index_f(t % TOTAL_CELLS, q)] = from[q * coarse_cells + d % coarse_cells];
-}
-
-// Experimental Schornbaum--Ruede (arXiv:1508.07982, Algorithm 2)
-// and Chen et al. (doi:10.1016/j.physa.2005.09.036) population interpolation.
-// Independent implementation of the mathematical stencil; no upstream source copied.
-// Stencil cells are owner cores; a disabled axis repeats its center twice.
-extern "C" __global__ void BlockExplodeProjected(
-    const fpxx* coarse, fpxx* fine, const uint64_t* target, const uint64_t* donor,
-    unsigned int count, unsigned long coarse_cells,
-    const uint64_t* neighbors, const float* offsets)
-{
-    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
-    if (row >= count) return;
-    const uint64_t t = target[row], d = donor[row];
-    fpxx* to = fine + (t / TOTAL_CELLS) * NQ * TOTAL_CELLS;
-    for (unsigned int q = 0; q < NQ; ++q) {
-        float v[4];
-        for (unsigned int a = 0; a < 4; ++a) {
-            const uint64_t n = neighbors[4ul * row + a];
-            v[a] = coarse[(n / coarse_cells) * NQ * coarse_cells
-                          + q * coarse_cells + n % coarse_cells];
-        }
-        float gx = .5f * (v[1] - v[0]), gy = .5f * (v[3] - v[2]);
-        const float ex = d_cx[q], ey = d_cy[q];
-        if (q != 0) {
-            const float along = (ex * gx + ey * gy) / (ex * ex + ey * ey);
-            gx -= ex * along;
-            gy -= ey * along;
-        }
-        to[index_f(t % TOTAL_CELLS, q)] =
-            coarse[(d / coarse_cells) * NQ * coarse_cells + q * coarse_cells + d % coarse_cells]
-            + offsets[2ul * row] * gx + offsets[2ul * row + 1] * gy;
-    }
-}
-
-// Targets are receiver-local SOURCE slots for the immediately following pull.
-// Child summation order ((c0 + c1) + c2) + c3 is part of the numerics; no collision here.
-extern "C" __global__ void BlockCoalesce(
-    const fpxx* fine, fpxx* coarse, const uint64_t* target, const uint64_t* children,
-    unsigned int count)
-{
-    unsigned int row = blockIdx.x * blockDim.x + threadIdx.x;
-    if (row >= count) return;
-    const uint64_t* s = children + 4ul * row;
-    coarse[target[row]] = 0.25f * (fine[s[0]] + fine[s[1]] + fine[s[2]] + fine[s[3]]);
-}
@@ -3,7 +3,7 @@
 //
 // Decodes the EsoPull backing layout into physical distributions, then
 // computes rho/ux/uy.  Output is SoA: rho_out[N], ux_out[N], uy_out[N].
-// Included by uniform.cu.  No standalone compilation.
+// Included by kernel_v2.cu.  No standalone compilation.
 //
 // This kernel is the sole path for macroscopic diagnostics in EsoPull mode.
 // It uses the shared semantic helper layer (esopull_semantic_helpers.cuh).
@@ -5,7 +5,7 @@
 // or encodes physical DDF back to EsoPull backing layout.
 // Each thread handles one cell from a compact index list.
 //
-// Included by uniform.cu.  No standalone compilation.
+// Included by kernel_v2.cu.  No standalone compilation.
 // ============================================================================

 #ifndef CELERIS_STEP_ESOPULL_SYNC_CU
@@ -1,6 +1,6 @@
 // CelerisLab – step/init_flow.cu
 // Channel-flow initialization kernels (double-buffer and EsoPull).
-// Included by uniform.cu after all headers.  No standalone compilation.
+// Included by kernel_v2.cu after all headers.  No standalone compilation.
 // ============================================================================

 #ifndef CELERIS_STEP_INIT_FLOW_CU
@@ -73,9 +73,9 @@ __device__ __forceinline__ uint16_t finalize_domain_flag(
 }

 // ============================================================================
-// InitTubeFlow — double-buffer channel initialization
+// InitTubeFlow_v2 — double-buffer channel initialization
 // ============================================================================
-__global__ void InitTubeFlow(uint16_t* flag, fpxx* fi)
+__global__ void InitTubeFlow_v2(uint16_t* flag, fpxx* fi)
 {
 #if DIM == 2
    unsigned int x, y; unsigned long k;
@@ -1,6 +1,6 @@
 // CelerisLab – step/one_step_double.cu
 // Double-buffer PULL streaming step kernel.
-// Included by uniform.cu.  No standalone compilation.
+// Included by kernel_v2.cu.  No standalone compilation.
 // ============================================================================

 #ifndef CELERIS_STEP_ONE_STEP_DOUBLE_CU
@@ -1,6 +1,6 @@
 // CelerisLab – step/one_step_esopull.cu
 // Esoteric-Pull single-buffer step kernel.
-// Included by uniform.cu.  No standalone compilation.
+// Included by kernel_v2.cu.  No standalone compilation.
 //
 // CURVED BC: supported via PrepareEsoPullCurvedKernel + ApplyEsoPullCurvedKernel
 // in aux_kernels.cu, launched by stepper.py before EsoPullStep.
@@ -21,14 +21,6 @@
 //   i.e. the population traveling in direction i at node n
 //   arrived from the node at n - c_i  = neighbor in direction opp(i)
 // ---------------------------------------------------------------------------
-// load_node: all NQ populations of node n as stored (no streaming).
-__device__ __forceinline__ void load_node(const fpxx* __restrict__ fi, unsigned long n,
-                                          float* __restrict__ f)
-{
-    for (int q = 0; q < NQ; ++q)
-        f[q] = load_ddf(fi, index_f(n, q));
-}
-
 __device__ __forceinline__ void stream_pull_load(unsigned long n,
                                                 float* __restrict__ f,
                                                 const fpxx* __restrict__ fi,
@@ -115,9 +115,8 @@ class FixedPatch2D:
        return self.bounds[0] - HALO / 2, self.bounds[1] - HALO / 2

    @property
-    def fine_level_offset(self) -> tuple[int, int]:
-        """Array index (y, x) of this window's first cell on the level-wide 2:1 fine grid
-        (an index offset, not an (x, y) origin)."""
+    def fine_level_origin(self) -> tuple[int, int]:
+        """(y, x) of this window's first cell on the level-wide 2:1 fine grid."""
        return 2 * self.bounds[1] - HALO, 2 * self.bounds[0] - HALO

    @property
@@ -152,6 +151,7 @@ class FixedPatch2D:
            raise ValueError("levels are not synchronized: fine_steps must equal 2*coarse_steps")


+MIN_PATCH_GAP = 4  # Legacy exported constant; union transfers impose no gap policy.


@dataclass(frozen=True, eq=False)
@@ -8,7 +8,7 @@ from dataclasses import dataclass, replace
 import numpy as np

 from .blocks import blocked_link_tables
-from ..body.jets import _surface_links
+from ..body.jets import jet_link_shapes, surface_link_velocity
 from .descriptors import D2Q9_EX, D2Q9_EY
 from .refinement import ACTIVE, OUTER_GHOST

@@ -36,7 +36,7 @@ def blocked_kind(layout, dense_kind):
    kind[owner[owned]] = np.asarray(dense_kind).ravel()[owned]
    return kind

-def blocked_coarse_tables(patch, layout, order, wall_aware=False, *, composed_kind=None):
+def blocked_base_tables(patch, layout, order, wall_aware=False, *, composed_kind=None):
    """Parent donor cells and receiver-local arrival slots, in transfer row order.

    Stage arrivals only after child reads and parent halo gather, immediately
@@ -142,13 +142,11 @@ def _prepare_blocked_boundary(boundary, layout, cfg, jets, *, _shifted_blocked=F
    tables.update(body_id=np.asarray(body_id, np.int32), rx=np.asarray(rx, np.float32),
                  ry=np.asarray(ry, np.float32), aperture=np.zeros(len(dirs), bool))
    actuated = bool(jets) or any(o.surface_actuation is not None for o in boundary.objects)
-    jet_cache = None
    if actuated:
        cfg = replace(cfg)
        cfg.validate()
-        tables["surface_velocity"], tables["aperture"], shape, jet_id = _surface_links(
+        tables["surface_velocity"], tables["aperture"] = surface_link_velocity(
            boundary.objects, cfg, boundary.flags, fluid_idx, dirs, body_id, rx, ry,
            jets=jets, hits=boundary.curved_hits, _shifted_blocked=_shifted_blocked)
-        if jets:
-            jet_cache = (shape, jet_id)
-    return _BlockedBoundary(boundary, tables, jets, actuated, jet_cache)
+    return _BlockedBoundary(boundary, tables, jets, actuated,
+                            jet_link_shapes(jets, boundary.curved_hits) if jets else None)
@@ -28,7 +28,7 @@ class LBMStepper:
            self.boundary_values_gpu = cuda.mem_alloc((6 * cfg.nx + 2 * cfg.ny * cfg.nq) * 4)
        else:
            self.step_fn = module.get_function("OneStep")
-            self.init_fn = module.get_function("InitTubeFlow")
+            self.init_fn = module.get_function("InitTubeFlow_v2")

        self.curved_fn = module.get_function("CurvedBoundaryKernel")
        self.prepare_esopull_curved_fn = module.get_function("PrepareEsoPullCurvedKernel")
@@ -25,7 +25,7 @@ import pycuda.driver as cuda

 from .config import LBMConfig, BodyConfig, load_lbm_config, load_body_config
 from .cuda.context import CudaContext, context_bound
-from .cuda import compiler
+from .cuda import compiler_v2 as compiler
 from .lbm.field import LBMField
 from .lbm.stepper import LBMStepper
 from .body.objects import SimObject
@@ -46,8 +46,8 @@ class Simulation:
                 device_id: int = 0):
        # Internal stream for async DTOH coordination
        self._stream: Optional[cuda.Stream] = None
+        self._build_dir = None
        self._closed = False
-        self._failed = None  # reason after an uncertain device write; only close() remains

        # Load configs
        self.lbm_cfg = load_lbm_config(lbm_config_path)
@@ -59,7 +59,9 @@ class Simulation:
        try:
            with self.ctx.activate():
                arch = self._resolve_compile_arch()
-                self._module = compiler.build_module(self.lbm_cfg, "uniform.cu", arch=arch)
+                self._build_dir = compiler.generate_config(self.lbm_cfg, n_objects=0)
+                self._ptx_path = compiler.compile_kernel(arch=arch, build_dir=self._build_dir)
+                self._module = compiler.load_module(self._ptx_path)
                self._compiled_n_objects = 0
                self._allocation_signature = self.lbm_cfg.allocation_signature()
                self._compiled_viscosity = self.lbm_cfg.viscosity
@@ -90,7 +92,6 @@ class Simulation:
    @context_bound
    def stream(self) -> cuda.Stream:
        """Internal CUDA stream for async DTOH coordination."""
-        self._require_usable()
        if self._stream is None:
            self._stream = cuda.Stream()
        return self._stream
@@ -174,27 +175,15 @@ class Simulation:
    def set_jet_amplitude(self, index: int, amplitude: float) -> None:
        """Change one jet's signed amplitude; re-uploads per-link velocity only.

-        Invalid input raises before any change. Any error after validation
-        (stream creation or upload) leaves the host amplitude unchanged but the
-        device column undefined, so the Simulation is marked failed: every later
-        device-touching public method (incl. ``stream``) raises; close() still
-        works. Close and recreate it.
+        Invalid input changes nothing. A device error leaves the host amplitude
+        unchanged but the device column undefined: treat it like any other CUDA
+        exception here and close/recreate the Simulation before stepping again.
        """
-        self._require_usable()
        jet = self.bodies.jets[index].with_amplitude(amplitude)
        if self._initialized:
-            try:
-                self.field.curved.set_jet_amplitude(index, amplitude, self.stream)
-            except BaseException as exc:
-                self._failed = f"jet amplitude upload failed: {exc!r}"
-                raise
+            self.field.curved.set_jet_amplitude(index, amplitude, self.stream)
        self.bodies.jets[index] = jet  # later geometry syncs keep the new amplitude

-    def _require_usable(self):
-        failed = getattr(self, "_failed", None)
-        if failed is not None:
-            raise RuntimeError(f"Simulation failed ({failed}); close() and recreate it")
-
    def remove_body(self, id: int) -> None:
        """Remove a body from the simulation.

@@ -352,7 +341,6 @@ class Simulation:
        Unlike read_body this is not accumulated/time-averaged. It resets after
        geometry sync or checkpoint load; advance before interpreting it.
        """
-        self._require_usable()
        obj = self.bodies.get(id)
        curved = self.field.curved
        has_aperture = (curved.aperture is not None
@@ -380,7 +368,6 @@ class Simulation:
            BodyTelemetry with fields ``force``, ``torque``, ``sensor``.
            ``sensor`` is a dim-length array of zeros for non-sensor bodies.
        """
-        self._require_usable()
        if self.bodies.obs_pinned is None:
            raise RuntimeError("No obs buffer. Call run() first.")
        if stream is None:
@@ -400,7 +387,6 @@ class Simulation:
        This is the most efficient readback for DRL control loops
        (single contiguous D2H copy, no Python per-body overhead).
        """
-        self._require_usable()
        if self.bodies.obs_pinned is None:
            raise RuntimeError("No obs buffer. Call run() first.")
        if stream is None:
@@ -426,8 +412,15 @@ class Simulation:
        self.lbm_cfg.validate_runtime()
        self._assert_allocation_contract()
        arch = self._resolve_compile_arch()
-        self._module = compiler.build_module(self.lbm_cfg, "uniform.cu", arch=arch,
-                                             n_objects=expected_n_objects)
+        next_dir = compiler.generate_config(self.lbm_cfg, n_objects=expected_n_objects)
+        try:
+            next_ptx = compiler.compile_kernel(arch=arch, build_dir=next_dir)
+            next_module = compiler.load_module(next_ptx)
+        except Exception:
+            compiler.cleanup_build_dir(next_dir)
+            raise
+        old_dir = self._build_dir
+        self._build_dir, self._ptx_path, self._module = next_dir, next_ptx, next_module
        self._compiled_n_objects = int(expected_n_objects)

        # Reconnect field and stepper to new module
@@ -445,6 +438,7 @@ class Simulation:
        self.stepper._step_count = _prev_step_count

        self._assert_object_count_contract(expected_count=expected_n_objects)
+        compiler.cleanup_build_dir(old_dir)

    def recompile(self):
        """Re-generate config headers and recompile kernel.
@@ -463,7 +457,6 @@ class Simulation:
    @context_bound
    def initialize(self):
        """Initialize flow field and sync objects to GPU."""
-        self._require_usable()
        self._assert_allocation_contract()
        # Validate the complete geometry before flags, populations or modules
        # change. Compact-list construction publishes no outputs on failure.
@@ -496,7 +489,6 @@ class Simulation:
        In EsoPull mode, the backing layout is decoded to physical DDF on
        GPU for the patched cells only, patched on host, then encoded back.
        """
-        self._require_usable()
        if not self._initialized:
            raise RuntimeError("Call initialize() before sync_bodies()")

@@ -592,7 +584,6 @@ class Simulation:
                sensor) on GPU and reset the step accumulator before stepping.
            checkpoint_interval: Nonnegative integer; if >0, save every N steps.
        """
-        self._require_usable()
        if not self._initialized:
            raise RuntimeError("Call initialize() first")
        steps = operator.index(steps)
@@ -657,12 +648,10 @@ class Simulation:
    @context_bound
    def get_macroscopic(self) -> Dict[str, np.ndarray]:
        """Download DDF and return rho, ux, uy [, uz]."""
-        self._require_usable()
        return self.field.get_macroscopic(step_id=self.stepper.step_count)

    @context_bound
    def get_ddf(self) -> np.ndarray:
-        self._require_usable()
        self.field.download_ddf(step_id=self.stepper.step_count)
        return self.field.ddf.copy()

@@ -676,27 +665,23 @@ class Simulation:

        Accepted: omega, fx, fy, fz, n_objects.
        """
-        self._require_usable()
        self.field.update_params(**kwargs)

    # -- Snapshots -----------------------------------------------------------
    @context_bound
    def snapshot(self):
        """Snapshot the flow field; restore keeps the current clock and controls."""
-        self._require_usable()
        self.field.snapshot(step_id=self.stepper.step_count)

    @context_bound
    def restore(self):
        """Restore the saved flow field at the current streaming parity."""
-        self._require_usable()
        self.field.restore(step_id=self.stepper.step_count)

    # -- Checkpoint / Restore (HDF5) ----------------------------------------
    @context_bound
    def save_checkpoint(self, path: str = None):
        """Save full simulation state to an HDF5 file."""
-        self._require_usable()
        from .common.checkpoint import save_checkpoint
        if self.bodies.jets:
            raise NotImplementedError("Jets are not checkpointed yet")
@@ -706,7 +691,6 @@ class Simulation:
    @context_bound
    def load_checkpoint(self, path: str):
        """Restore simulation state from an HDF5 checkpoint."""
-        self._require_usable()
        from .common.checkpoint import load_checkpoint
        if not self._initialized:
            self.initialize()
@@ -731,6 +715,8 @@ class Simulation:
                self._module = None
                self._stream = None
            ctx.close()
+        compiler.cleanup_build_dir(getattr(self, "_build_dir", None))
+        self._build_dir = None
        self._closed = True

    def __enter__(self):
@@ -1,66 +0,0 @@
-"""Checkpoint save -> load round trip for a scene with passive bodies. Requires GPU.
-
-A fresh Simulation with the same scene, restored from the checkpoint must hold the same populations,
-flags, body count and step count, and must stay bitwise identical to the original
-after both continue for the same number of steps. Checkpoints with Jets are
-refused (documented limitation) and are not covered here.
-"""
-
-import os
-import tempfile
-import unittest
-
-import numpy as np
-
-from ._cases import make_sim
-
-# Three separated cylinders (pinball-like); cut-link receivers stay in bulk fluid.
-CIRCLES = (((40, 32), 6), ((56, 22), 6), ((56, 42), 6))
-
-
-def scene():
-    sim = make_sim()
-    sim.initialize()
-    for center, radius in CIRCLES:
-        sim.add_body("circle", center=center, radius=radius)
-    sim.sync_bodies()
-    return sim
-
-
-class CheckpointBodiesGPU(unittest.TestCase):
-    def test_round_trip_is_bitwise_and_continues_identically(self):
-        sim = scene()
-        sim.run(60)
-        with tempfile.NamedTemporaryFile(suffix=".h5", delete=False) as f:
-            path = f.name
-        restored = None
-        try:
-            saved = sim.save_checkpoint(path)
-            # The loader requires the same body count and geometry: rebuild the scene.
-            restored = scene()
-            restored.load_checkpoint(saved)
-
-            self.assertEqual(restored.bodies.count, len(CIRCLES))
-            self.assertEqual(restored.stepper.step_count, sim.stepper.step_count)
-            np.testing.assert_array_equal(restored.get_flags(), sim.get_flags())
-            a, b = sim.get_ddf(), restored.get_ddf()
-            self.assertEqual(a.dtype, b.dtype)
-            self.assertEqual(a.tobytes(), b.tobytes())
-
-            sim.run(40)
-            restored.run(40)
-            self.assertEqual(restored.stepper.step_count, sim.stepper.step_count)
-            a, b = sim.get_ddf(), restored.get_ddf()
-            self.assertTrue(np.isfinite(a).all())
-            self.assertEqual(a.tobytes(), b.tobytes())
-            for i in range(len(CIRCLES)):
-                self.assertTrue(np.all(np.isfinite(restored.read_force(i))))
-        finally:
-            if restored is not None:
-                restored.close()
-            os.unlink(path)
-            sim.close()
-
-
-if __name__ == "__main__":
-    unittest.main()
@@ -1,53 +0,0 @@
-"""Direct Simulation Jet control upload; boundary-law accuracy is tested elsewhere."""
-import numpy as np
-import pycuda.driver as cuda
-
-from CelerisLab.body.jets import Jet, Radial, Sector
-from CelerisLab.lbm.descriptors import FLUID
-from ._cases import make_sim
-
-
-def test_simulation_jet_amplitude_upload_and_continued_steps():
-    # Same grid/circle/sector as the existing uniform surface-actuation fixture.
-    center = (16., 12.)
-    with make_sim(nx=32, ny=24, velocity=0) as sim:
-        sim.add_body("circle", center=center, radius=4.3)
-        index = sim.add_jet(Jet(Sector(center, .7, .85, 4.4),
-                                Radial(center, inward=False), "quartic", .002))
-        sim.initialize()
-        curved = sim.field.curved
-        shape, jet_id = curved.jet_shape.copy(), curved.jet_id.copy()
-        flags = sim.get_flags()
-        driven = jet_id == index
-        assert driven.any() and (~driven).any()
-        velocity_pointer = int(curved.surface_velocity_gpu)
-        previous = None
-        for interval, amplitude in enumerate((.002, -.001)):
-            if interval:
-                count = sim.stepper.step_count
-                sim.set_jet_amplitude(index, amplitude)
-                assert sim.stepper.step_count == count  # update does not step
-            expected = np.zeros_like(curved.surface_velocity)
-            expected[driven] = (amplitude * shape[driven]).astype(np.float32)
-            actual = np.empty_like(expected)
-            with sim.ctx.activate():
-                cuda.memcpy_dtoh(actual, curved.surface_velocity_gpu)
-            np.testing.assert_array_equal(actual, expected)  # actual device column
-            np.testing.assert_array_equal(curved.surface_velocity, expected)
-            assert np.any(actual[driven] != 0)
-            if previous is not None:
-                assert not np.array_equal(actual[driven], previous[driven])
-            previous = actual.copy()
-            assert sim.bodies.jets[index].amplitude == amplitude
-            assert curved.jets[index].amplitude == amplitude
-            assert int(curved.surface_velocity_gpu) == velocity_pointer
-            np.testing.assert_array_equal(curved.jet_shape, shape)
-            np.testing.assert_array_equal(curved.jet_id, jet_id)
-            np.testing.assert_array_equal(sim.get_flags(), flags)
-            sim.run(2)
-            assert sim.stepper.step_count == 2 * (interval + 1)
-            assert sim._failed is None
-            macro = sim.get_macroscopic()
-            fluid = (flags.reshape(24, 32) & FLUID) != 0
-            assert all(np.isfinite(value[fluid]).all() for value in macro.values())
-            assert np.all(macro["rho"][fluid] > 0)
@@ -6,14 +6,14 @@ import pycuda.driver as cuda
 import pytest

 import CelerisLab
-from CelerisLab.cuda import compiler
+from CelerisLab.cuda import compiler_v2
 from ._cases import make_sim


 def test_source_provenance():
    root = Path(__file__).resolve().parents[2] / "src/CelerisLab"
    assert Path(CelerisLab.__file__).resolve().parent == root
-    assert Path(compiler.kernel_path("uniform.cu")).resolve() == root / "lbm/kernels/uniform.cu"
+    assert Path(compiler_v2.kernel_path("kernel_v2.cu")).resolve() == root / "lbm/kernels/kernel_v2.cu"


 def test_two_contexts_interleaved_and_non_lifo_close():
@@ -21,7 +21,7 @@ def test_two_contexts_interleaved_and_non_lifo_close():
    first = make_sim(nx=32, ny=16)
    second = make_sim(nx=48, ny=24, streaming="esopull", shifted=True)
    try:
-        assert first._module is not second._module
+        assert first._build_dir != second._build_dir
        assert cuda.Context.get_current() == before
        first.initialize()
        second.initialize()
@@ -16,7 +16,6 @@
 #define NZ 1
 #define NT 1
 #define COLLISION_MODEL 0
-#define BLOCK_ROOT_BC 2
 #define USE_LES 0
 #define RHO 1.f
 #define U0 .025f
@@ -41,7 +40,7 @@ inline float __fdiv_rn(float a,float b){return a/b;}
 #include "boundary/inlet_outlet.cuh"
 #include "step/one_step_double.cu"
 #else
-#include "core/cell_kind.cuh"
+constexpr unsigned char KIND_OUTER=2;
 #include "step/block_channel.cu"
 #endif
 int main(){
@@ -1,46 +0,0 @@
-"""Check historical source manifests against their own recorded commit, not today's checkout."""
-
-import io
-from pathlib import Path
-import subprocess
-import tarfile
-
-ROOT = Path(__file__).resolve().parents[2]
-
-
-def _require(commit):
-    """Missing recorded history is an explicit failure, never a skip or a repin."""
-    found = subprocess.run(
-        ["git", "-C", str(ROOT), "cat-file", "-e", commit + "^{commit}"],
-        capture_output=True,
-    )
-    if found.returncode:
-        raise AssertionError(
-            f"Required history missing: commit {commit} is not in this clone; "
-            "fetch it instead of repinning or substituting"
-        )
-
-
-def tracked(commit, prefix="src"):
-    """Tracked paths under prefix at the recorded commit."""
-    _require(commit)
-    return set(
-        subprocess.check_output(
-            ["git", "-C", str(ROOT), "ls-tree", "-r", "--name-only", commit, prefix],
-            text=True,
-        ).splitlines()
-    )
-
-
-def historical_tree(commit, dest):
-    """Extract the tracked tree of the recorded commit into dest."""
-    _require(commit)
-    archive = subprocess.check_output(
-        ["git", "-C", str(ROOT), "archive", "--format=tar", commit]
-    )
-    dest = Path(dest)
-    dest.mkdir(parents=True, exist_ok=True)
-    extra = {"filter": "data"} if hasattr(tarfile, "data_filter") else {}
-    with tarfile.open(fileobj=io.BytesIO(archive)) as tar:
-        tar.extractall(dest, **extra)
-    return dest
@@ -0,0 +1,101 @@
+"""Synthetic checks of the annulus mass ledger. Host-only, stdlib; not solver evidence."""
+
+import random
+import sys
+from pathlib import Path
+import unittest
+
+sys.path.insert(0, str(Path(__file__).resolve().parents[1] / "validation"))
+import annulus_ledger as al  # noqa: E402
+
+NX = NY = 24
+# Annulus 4 <= r <= 8 around (12, 12); dyadic values keep every sum exact.
+A = {(x, y) for x in range(NX) for y in range(NY) if 16 <= (x - 12) ** 2 + (y - 12) ** 2 <= 64}
+DELTA = 0.375
+
+
+def owner(seed=1):
+    rng = random.Random(seed)
+    return [[[rng.randint(1, 1024) / 1024 for _ in range(NX)] for _ in range(NY)] for _ in range(9)]
+
+
+def collide(P, source=None):
+    """Mass-conserving stand-in: rotate populations within each cell (+ optional source)."""
+    D = [[[P[(q + 1) % 9][y][x] for x in range(NX)] for y in range(NY)] for q in range(9)]
+    if source:
+        (x, y), s = source
+        D[0][y][x] += s
+    return D
+
+
+def link(interior):
+    """(x, y, q) with x in A and x - c_q inside (seam) or outside (edge) A."""
+    for (x, y) in sorted(A):
+        for q in range(1, 9):
+            src = ((x - al.CX[q]) % NX, (y - al.CY[q]) % NY)
+            if (src in A) == interior:
+                return x, y, q
+    raise AssertionError
+
+
+class Ledger(unittest.TestCase):
+    def setUp(self):
+        self.O = owner()
+        self.P = al.pull(self.O, NX, NY)
+
+    def run_ledger(self, P, D):
+        return al.ledger(self.O, P, D, A, NX, NY)
+
+    def test_ideal_pull_has_no_h_and_closes(self):
+        r = self.run_ledger(self.P, collide(self.P))
+        self.assertEqual((r["H_seam"], r["H_edge"], r["C"], r["residual"]), (0, 0, 0, 0))
+        self.assertEqual(r["dM"], -r["Q_exact"])
+        self.assertNotEqual(r["Q_exact"], 0)
+
+    def test_collision_source_is_c(self):
+        x, y = sorted(A)[3]
+        r = self.run_ledger(self.P, collide(self.P, ((x, y), DELTA)))
+        self.assertEqual((r["C"], r["H"], r["residual"]), (DELTA, 0, 0))
+
+    def test_interior_seam_perturbation(self):
+        x, y, q = link(interior=True)
+        self.P[q][y][x] += DELTA
+        r = self.run_ledger(self.P, collide(self.P))
+        self.assertEqual((r["H_seam"], r["H_edge"], r["residual"]), (DELTA, 0, 0))
+        # Invisible to any flux through the edge of A; only H explains the mass change.
+        self.assertEqual(al.staged_flux(self.O, self.P, A, NX, NY), r["Q_exact"])
+        self.assertEqual(r["dM"] + r["Q_exact"], DELTA)
+
+    def test_edge_perturbation_and_staged_flux_double_count(self):
+        x, y, q = link(interior=False)
+        self.P[q][y][x] += DELTA
+        r = self.run_ledger(self.P, collide(self.P))
+        self.assertEqual((r["H_seam"], r["H_edge"], r["residual"]), (0, DELTA, 0))
+        staged = al.staged_flux(self.O, self.P, A, NX, NY)
+        self.assertEqual(staged, r["Q_exact"] - DELTA)
+        # Staged flux + full H counts the staging change twice.
+        self.assertEqual(r["dM"] + staged - r["C"] - r["H"], -DELTA)
+        # Staged flux alone closes only without the edge part of H.
+        self.assertEqual(r["dM"] + staged - r["C"] - r["H_seam"], 0)
+
+    def test_owner_snapshot_after_staging_hides_the_change(self):
+        # Snapshot taken after staging wrote the same slot into the owner image.
+        x, y, q = link(interior=False)
+        sx, sy = (x - al.CX[q]) % NX, (y - al.CY[q]) % NY
+        self.P[q][y][x] += DELTA
+        late = [[row[:] for row in plane] for plane in self.O]
+        late[q][sy][sx] += DELTA
+        r = al.ledger(late, self.P, collide(self.P), A, NX, NY)
+        self.assertEqual((r["H"], r["residual"]), (0, 0))
+
+    def test_wrong_pull_direction_is_flagged(self):
+        # Indexing error (push instead of pull): H is nonzero with no physical cause.
+        bad = [[[self.O[q][(y + al.CY[q]) % NY][(x + al.CX[q]) % NX] for x in range(NX)]
+                for y in range(NY)] for q in range(9)]
+        r = self.run_ledger(bad, collide(bad))
+        self.assertNotEqual(r["H"], 0)
+        self.assertEqual(r["residual"], 0)
+
+
+if __name__ == "__main__":
+    unittest.main()
@@ -83,32 +83,6 @@ assert abs(sum(compute_feq_d2q9(1.,0.,0.)) - 1.) < 1e-6


 class BuildContracts(unittest.TestCase):
-    def test_block_root_modes_are_rendered_and_recorded(self):
-        with tempfile.TemporaryDirectory() as root:
-            for mode in (0, 1, 2):
-                with self.subTest(mode=mode):
-                    build = compiler.generate_config(LBMConfig(), build_root=root, block_root_bc=mode)
-                    try:
-                        method = Path(build, "config/config_method.h").read_text()
-                        self.assertIn(f"#define BLOCK_ROOT_BC {mode}\n", method)
-                        metadata = json.loads(Path(build, ".celeris-build").read_text())
-                        self.assertEqual(metadata["block_root_bc"], mode)
-                        self.assertNotIn("block_root_bc", metadata["config"])
-                    finally:
-                        compiler.cleanup_build_dir(build)
-            build = compiler.generate_config(LBMConfig(), build_root=root)
-            try:
-                self.assertEqual(json.loads(Path(build, ".celeris-build").read_text())["block_root_bc"], 0)
-            finally:
-                compiler.cleanup_build_dir(build)
-
-    def test_invalid_block_root_mode_fails_before_snapshot(self):
-        for mode in (-1, 3, True, False, 1.0, "1", None):
-            with self.subTest(mode=mode), patch.object(compiler.tempfile, "mkdtemp") as allocate:
-                with self.assertRaisesRegex(ValueError, "block_root_bc"):
-                    compiler.generate_config(LBMConfig(), block_root_bc=mode)
-                allocate.assert_not_called()
-
    @staticmethod
    def fake_compile(command, check):
        source = Path(command[-3])
@@ -126,7 +100,8 @@ class BuildContracts(unittest.TestCase):
            def work(nx, count):
                build = compiler.generate_config(LBMConfig(nx=nx), count, build_root=root)
                barrier.wait()
-                output = compiler.compile_kernel(build_dir=build)
+                # Compatibility form is context-local even without explicit path.
+                output = compiler.compile_kernel()
                data = json.loads(Path(output).read_text())
                self.assertEqual(data, {"nx": nx, "objects": count})
                metadata = json.loads(Path(output + ".json").read_text())
@@ -155,7 +130,7 @@ class BuildContracts(unittest.TestCase):
            compiler.cleanup_build_dir(a)
            compiler.cleanup_build_dir(b)
            with self.assertRaises(RuntimeError):
-                compiler.compile_kernel(build_dir=a)  # removed snapshot is never compiled
+                compiler.compile_kernel()

    def test_failed_compile_cannot_implicitly_load_previous_artifact(self):
        with tempfile.TemporaryDirectory() as root:
@@ -168,7 +143,8 @@ class BuildContracts(unittest.TestCase):
                with self.assertRaises(subprocess.CalledProcessError):
                    compiler.compile_kernel(build_dir=build)
            self.assertEqual(Path(previous).read_bytes(), old_bytes)
-            self.assertEqual(sorted(Path(build).glob("uniform-*.ptx")), [])  # no partial output
+            with self.assertRaisesRegex(RuntimeError, "No compiled PTX"):
+                compiler.load_module()
            compiler.cleanup_build_dir(build)


@@ -7,7 +7,7 @@ from CelerisLab.config import LBMConfig
 from CelerisLab.lbm.blocks import BlockLayout2D
 from CelerisLab.lbm.channel import channel_masks, validate_channel_config, validate_channel_transfers
 from CelerisLab.lbm.refinement import FixedPatch2D
-from CelerisLab.lbm.refinement_tables import blocked_coarse_tables
+from CelerisLab.lbm.refinement_tables import blocked_base_tables
 from .test_numerical_operators import compile_run


@@ -31,7 +31,7 @@ def transfer(bounds, block=(4, 6)):
    cfg = LBMConfig(nx=24, ny=20, viscosity=.1)
    pair = FixedPatch2D.from_config(cfg, bounds)
    layout = BlockLayout2D.build(20, 24, *block)
-    targets = blocked_coarse_tables(pair, layout, np.argsort(pair.coalesce_coarse))[1]
+    targets = blocked_base_tables(pair, layout, np.argsort(pair.coalesce_coarse))[1]
    return pair, layout, targets


@@ -50,7 +50,7 @@ def test_masks_and_minimal_footprint():
 def test_actual_closure_read_conflicts():
    pair, layout, targets = transfer((6, 6, 18, 14))
    validate_channel_transfers(pair, layout, targets)
-    for target in (layout.slot(int(layout.owner[5, -1]) - 1, 2),  # integer under NumPy 1 and 2
+    for target in (layout.slot(layout.owner[5, -1]-1, 2),
                   layout.slot(layout.owner[1, 4], 4),
                   layout.slot(layout.owner[-2, 4], 3)):
        with pytest.raises(ValueError, match="immutable"):
@@ -74,14 +74,14 @@ def test_evolved_ghost_snapshot_and_physical_download():
    from .test_refinement_three_level_multipatch import host_module
    host = host_module()
    storage = {}
-    host.compiler.build_module.return_value.get_global.side_effect = lambda name: (name, 12 if name == "d_block_channel" else 20)
+    host.compiler.load_module.return_value.get_global.side_effect = lambda name: (name, 12 if name == "d_block_channel" else 20)
    host.cuda.mem_alloc.side_effect = lambda size: Mock()
    host.cuda.memcpy_htod.side_effect = lambda ptr, a: storage.__setitem__(id(ptr), np.array(a, copy=True))
    host.cuda.memcpy_dtoh.side_effect = lambda a, ptr: np.copyto(a, storage[id(ptr)])
    context = Mock(activate=nullcontext, sm_arch="sm_70")
    cfg = LBMConfig(nx=24, ny=20, viscosity=.1, velocity=.025)
-    backend = host.FixedPatchStepper(cfg, context=context, coarse_boundary="channel",
-                                    refinements=[(6, 6, 18, 14)], level_blocks=[(4, 6), (4, 4)])
+    backend = host.FixedPatchStepper(cfg, (6, 6, 18, 14), context=context,
+                                    coarse_boundary="channel", base_block=(4, 6), fine_block=(4, 4))
    try:
        weights = np.array([4/9]+[1/9]*4+[1/36]*4)[:, None, None]
        arrays = [np.broadcast_to(weights, level.shape).copy() for level in backend._levels]
@@ -96,7 +96,6 @@ def test_evolved_ghost_snapshot_and_physical_download():
        arrays[0][0, 0, 0] = np.nan
        with pytest.raises(ValueError, match="finite"):
            backend.initialize(*arrays)
-        assert backend._levels[0].kernel == host.compiler.build_module.return_value.get_function("BlockChannelStep")
-        assert [call.kwargs["block_root_bc"] for call in host.compiler.build_module.call_args_list] == [2, 0]
+        assert backend._levels[0].kernel == host.compiler.load_module.return_value.get_function("BlockChannelStep")
    finally:
        backend.close()
@@ -1,70 +0,0 @@
-"""The public examples run all-blocked: their level tiles must reach the device unchanged.
-
-Host only: a context whose activation raises proves construction (layouts, tables,
-boundary checks) succeeded up to the first CUDA call.
-"""
-import ast
-from dataclasses import replace
-from pathlib import Path
-import unittest
-from unittest.mock import Mock
-
-from CelerisLab.body.geometry.circle import CircleGeometry
-from CelerisLab.body.objects import SimObject
-from CelerisLab.config import LBMConfig
-from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
-
-EXAMPLES = Path(__file__).resolve().parents[2] / "examples"
-
-
-def literal(name, constant):
-    """Module-level literal from an example, read without importing it (no pycuda)."""
-    tree = ast.parse((EXAMPLES / name).read_text())
-    for node in tree.body:
-        if isinstance(node, ast.Assign):
-            for target in node.targets:
-                names = target.elts if isinstance(target, ast.Tuple) else [target]
-                for i, t in enumerate(names):
-                    if isinstance(t, ast.Name) and t.id == constant:
-                        value = node.value.elts[i] if isinstance(target, ast.Tuple) else node.value
-                        return ast.literal_eval(value)
-    raise KeyError(constant)
-
-
-def forbidden():
-    return Mock(activate=Mock(side_effect=AssertionError("CUDA reached")), sm_arch="sm_70")
-
-
-class ExampleTiles(unittest.TestCase):
-    def test_fixed_patch_jet_tiles(self):
-        bounds, blocks = literal("fixed_patch_jet.py", "BOUNDS"), literal("fixed_patch_jet.py", "LEVEL_BLOCKS")
-        width, amplitude = literal("fixed_patch_jet.py", "WIDTH"), literal("fixed_patch_jet.py", "AMPLITUDE")
-        cfg = LBMConfig(nx=48, ny=48, viscosity=.1, velocity=0., wall_slot_width=width,
-                        wall_normal_velocity=amplitude, y_wall_bc="on_site_velocity",
-                        top_wall_bc="reservoir")
-        with self.assertRaisesRegex(AssertionError, "CUDA reached"):
-            FixedPatchStepper(cfg, context=forbidden(), coarse_boundary="slot_reservoir",
-                              refinements=[bounds], level_blocks=[tuple(b) for b in blocks])
-
-    def test_fixed_periodic_patch_tiles(self):
-        with self.assertRaisesRegex(AssertionError, "CUDA reached"):
-            FixedPatchStepper(LBMConfig(nx=16, ny=16, viscosity=.1, velocity=1e-3), context=forbidden(),
-                              refinements=[(4, 4, 12, 12)], level_blocks=[(16, 16), (24, 24)])
-
-    def test_fixed_patch_circle_default_tiles(self):
-        bounds = literal("fixed_patch_circle.py", "BOUNDS")
-        nx, ny = literal("fixed_patch_circle.py", "NX"), literal("fixed_patch_circle.py", "NY")
-        center, radius = literal("fixed_patch_circle.py", "CENTER"), literal("fixed_patch_circle.py", "RADIUS")
-        body = SimObject(0, geometry=CircleGeometry(*center, radius), center=center, radius=radius)
-        periodic = LBMConfig(nx=nx, ny=ny, viscosity=.1)
-        slot = replace(periodic, velocity=0, wall_normal_velocity=.001, wall_slot_width=8,
-                       y_wall_bc="on_site_velocity", top_wall_bc="reservoir")
-        for mode, cfg in (("periodic", periodic), ("slot_reservoir", slot)):
-            with self.subTest(mode=mode), self.assertRaisesRegex(AssertionError, "CUDA reached"):
-                FixedPatchStepper(cfg, context=forbidden(), coarse_boundary=mode,
-                                  refinements=[bounds], level_blocks=[(12, 8), (16, 20)],
-                                  level_objects=[(), (body,)])
-
-
-if __name__ == "__main__":
-    unittest.main()
@@ -8,7 +8,6 @@ import numpy as np
 ROOT = Path(__file__).resolve().parents[2]
 REPO = ROOT
 sys.path.insert(0, str(ROOT / "tests/validation"))
-sys.path.insert(0, str(ROOT / "tests/unit"))
 sys.path.insert(0, str(REPO / "src"))
 from run_jet_circle_cv import (
    CX,
@@ -20,7 +19,6 @@ from run_jet_circle_cv import (
    cv_from_macros,
    aperture_probe,
    measured_aperture_flux,
-    PIN,
    verify,
 )

@@ -168,13 +166,7 @@ class Host(unittest.TestCase):
        self.assertEqual(a["steps"] * 4, b["steps"])

    def test_source_pin(self):
-        from tempfile import TemporaryDirectory
-        from historical_tree import historical_tree
-
-        with self.assertRaisesRegex(RuntimeError, "Source hash mismatch"):
-            verify(REPO)
-        with TemporaryDirectory() as directory:
-            verify(historical_tree(PIN, directory))
+        verify(REPO)


 if __name__ == "__main__":
@@ -1,149 +0,0 @@
-"""Jet amplitude update failure contract (host doubles only, no CUDA numerics).
-
-Both owners validate before any device write, commit host state only after the
-upload returns, and refuse every later call except close() after an uncertain
-device write.
-"""
-from types import SimpleNamespace
-
-import numpy as np
-import pytest
-
-pytest.importorskip("pycuda.driver")
-from CelerisLab import simulation as module
-from CelerisLab.body.jets import Jet, Radial, Sector
-from tests.unit.test_pinned_body_remap import ready
-from tests.unit.test_simulation_lifecycle import make_simulation
-
-BAD = (float("nan"), float("inf"), 1.)
-
-
-class Curved:
-    """Records uploads; ``fail`` raises once the write was attempted."""
-    def __init__(self):
-        self.writes, self.fail = [], False
-
-    def set_jet_amplitude(self, index, amplitude, stream):
-        self.writes.append((index, amplitude))
-        if self.fail:
-            raise RuntimeError("upload failed")
-
-
-def simulation(monkeypatch):
-    sim = make_simulation(monkeypatch)
-    jet = Jet(Sector((6., 5.), 0., .5, 4.), Radial((6., 5.)), "uniform", 1e-3)
-    sim._failed, sim._initialized = None, True
-    sim.bodies, sim.field = SimpleNamespace(jets=[jet]), SimpleNamespace(curved=Curved())
-    return sim, jet
-
-
-def test_simulation_invalid_input_writes_nothing_and_stays_usable(monkeypatch):
-    sim, jet = simulation(monkeypatch)
-    for amplitude in BAD:
-        with pytest.raises(ValueError):
-            sim.set_jet_amplitude(0, amplitude)
-    with pytest.raises(IndexError):
-        sim.set_jet_amplitude(1, 2e-3)
-    assert sim.field.curved.writes == [] and sim.bodies.jets[0] is jet and sim._failed is None
-    sim.set_jet_amplitude(0, 2e-3)  # retry after a rejected input
-    assert sim.field.curved.writes == [(0, 2e-3)] and sim.bodies.jets[0].amplitude == 2e-3
-    sim.close()
-
-
-def test_simulation_upload_failure_keeps_host_and_blocks_reuse(monkeypatch):
-    sim, jet = simulation(monkeypatch)
-    sim.field.curved.fail = True
-    with pytest.raises(RuntimeError, match="upload failed"):
-        sim.set_jet_amplitude(0, 2e-3)
-    assert sim.bodies.jets[0] is jet and sim._failed is not None
-    sim.field.curved.fail = False
-    for call in (lambda: sim.set_jet_amplitude(0, 2e-3), lambda: sim.run(1), sim.step,
-                 sim.sync_bodies, sim.initialize, sim.get_ddf, sim.get_macroscopic,
-                 lambda: sim.read_surface_exchange(0), lambda: sim.read_body(0),
-                 sim.read_bodies, sim.snapshot, sim.restore, sim.save_checkpoint,
-                 lambda: sim.load_checkpoint("x.h5"), lambda: sim.update_runtime_params(omega=1.)):
-        with pytest.raises(RuntimeError, match="Simulation failed"):
-            call()
-    assert sim.field.curved.writes == [(0, 2e-3)]
-    sim.close()  # cleanup still works
-    assert sim._closed and sim.ctx.is_closed
-
-
-def test_simulation_stream_creation_failure_fails_closed(monkeypatch):
-    sim, jet = simulation(monkeypatch)
-    created = []
-    def stream():
-        created.append(1)
-        raise RuntimeError("stream creation failed")
-    monkeypatch.setattr(module.cuda, "Stream", stream, raising=False)
-    sim._stream = None
-    with pytest.raises(RuntimeError, match="stream creation failed"):
-        sim.set_jet_amplitude(0, 2e-3)
-    assert sim.bodies.jets[0] is jet and sim._failed is not None
-    assert sim.field.curved.writes == [] and created == [1]
-    for call in (lambda: sim.stream, lambda: sim.set_jet_amplitude(0, 2e-3)):
-        with pytest.raises(RuntimeError, match="Simulation failed"):
-            call()
-    assert created == [1]  # the public stream property does not retry creation
-    sim.ctx.detach_failure = True
-    with pytest.raises(RuntimeError, match="detach failed"):
-        sim.close()
-    assert not sim._closed
-    sim.close()  # close stays retryable after the failure
-    assert sim._closed and sim.ctx.is_closed
-
-
-def test_simulation_before_initialize_commits_host_only(monkeypatch):
-    sim, _ = simulation(monkeypatch)
-    sim._initialized = False
-    sim.set_jet_amplitude(0, -2e-3)
-    assert sim.field.curved.writes == [] and sim.bodies.jets[0].amplitude == -2e-3
-    sim.close()
-
-
-def stepper_state(run):
-    boundary = run._device_boundaries[2]
-    return boundary.tables["surface_velocity"].copy(), boundary.jets, run._level_jets[2]
-
-
-def test_stepper_invalid_input_writes_nothing_and_stays_ready():
-    run, module, _, _, _ = ready()
-    try:
-        velocity, jets, level_jets = stepper_state(run)
-        module.cuda.memcpy_htod.reset_mock()
-        for args, kw, error in [((0, a), {"level": 2}, ValueError) for a in BAD] + [
-                ((1, .01), {"level": 2}, IndexError), ((0, .01), {"level": 3}, ValueError)]:
-            with pytest.raises(error):
-                run.set_jet_amplitude(*args, **kw)
-        module.cuda.memcpy_htod.assert_not_called()
-        np.testing.assert_array_equal(run._device_boundaries[2].tables["surface_velocity"], velocity)
-        assert (run._device_boundaries[2].jets, run._level_jets[2]) == (jets, level_jets)
-        assert run._status == "ready"
-        run.set_jet_amplitude(0, .017, level=2)
-        assert run._level_jets[2][0].amplitude == .017
-    finally:
-        run.close()
-
-
-@pytest.mark.parametrize("where", ["upload", "synchronize"])
-def test_stepper_device_failure_keeps_host_and_blocks_reuse(where):
-    run, module, _, _, _ = ready()
-    try:
-        velocity, jets, level_jets = stepper_state(run)
-        if where == "upload":
-            module.cuda.memcpy_htod.side_effect = RuntimeError("upload failed")
-        else:
-            run._stream.synchronize.side_effect = RuntimeError("upload failed")
-        with pytest.raises(RuntimeError, match="upload failed"):
-            run.set_jet_amplitude(0, .017, level=2)
-        np.testing.assert_array_equal(run._device_boundaries[2].tables["surface_velocity"], velocity)
-        assert (run._device_boundaries[2].jets, run._level_jets[2]) == (jets, level_jets)
-        assert run._status == "failed"
-        module.cuda.memcpy_htod.side_effect = None
-        run._stream.synchronize.side_effect = None
-        for call in (lambda: run.set_jet_amplitude(0, .017, level=2), run.step, run.download):
-            with pytest.raises(RuntimeError, match="failed"):
-                call()
-    finally:
-        run.close()
-    assert run._status == "closed"
@@ -194,7 +194,6 @@ def test_public_surface_exchange_for_jet_preserves_aperture_at_zero_and_reverse(
    field, bodies = jet_case(device, [jet])
    sim = SimpleNamespace(bodies=bodies, field=field, _initialized=True,
                          stream=SimpleNamespace(synchronize=device.synchronize))
-    sim._require_usable = Simulation._require_usable.__get__(sim)
    cl = field.curved
    assert bodies.get(0).surface_actuation is None
    aperture, velocity = cl.aperture, cl.surface_velocity.copy()
@@ -219,7 +218,6 @@ def test_public_surface_exchange_keeps_legacy_and_passive_guards(device):
    bodies.sync_to_gpu(field)
    sim = SimpleNamespace(bodies=bodies, field=field,
                          stream=SimpleNamespace(synchronize=device.synchronize))
-    sim._require_usable = Simulation._require_usable.__get__(sim)
    for value in Simulation.read_surface_exchange(sim, 0).values():
        np.testing.assert_array_equal(value, 0.)
    field.curved.free_gpu_columns()
@@ -15,10 +15,6 @@ sys.path.insert(0, str(Path(__file__).resolve().parents[1] / "validation"))
 from run_uniform_les_flow import source_guard, decode_macros, HERE, HEADER

 ROOT = HERE.parents[1]
-sys.path.insert(0, str(ROOT / "tests/unit"))
-from historical_tree import historical_tree, tracked
-
-A135 = "a135252ae78b47b57d7e77d19001700b3bedac8b"


 class GuardTests(unittest.TestCase):
@@ -46,23 +42,19 @@ class GuardTests(unittest.TestCase):
            decode_macros(invalid, w, cx, cy)

    def test_current_corrected_pins(self):
-        with TemporaryDirectory() as directory:
-            root = historical_tree(A135, directory)
-            with patch("run_uniform_les_flow.subprocess.check_output", return_value="historical-tree\n"):
-                identity = source_guard(root, "corrected", "a135252")
-            # base_commit comes from the pins file; the mock only feeds checkout_commit.
-            self.assertEqual(identity["base_commit"], A135)
-            self.assertEqual(identity["checkout_commit"], "historical-tree")
+        self.assertEqual(
+            source_guard(ROOT, "corrected", "a135252")["base_commit"],
+            "a135252ae78b47b57d7e77d19001700b3bedac8b",
+        )

    def test_corrected_and_tamper_guard(self):
        pins = json.loads((HERE / "data/les_flow/source_pins_a135252.json").read_text())
        with TemporaryDirectory() as directory:
-            origin = historical_tree(A135, Path(directory) / "origin")
-            root = Path(directory) / "fixture"
+            root = Path(directory)
            for name in pins["sources"]:
                dest = root / name
                dest.parent.mkdir(parents=True, exist_ok=True)
-                shutil.copyfile(origin / name, dest)
+                shutil.copyfile(ROOT / name, dest)
            with patch(
                "run_uniform_les_flow.subprocess.check_output",
                return_value="isolated-fixture\n",
@@ -95,26 +87,23 @@ class GuardTests(unittest.TestCase):

    def test_current_manifest_pins_actual_corrected_tree(self):
        pins = json.loads((HERE / "data/les_flow/source_pins_a135252.json").read_text())
-        self.assertEqual(pins["base_commit"], A135)
-        self.assertEqual(set(pins["sources"]), tracked(A135))
-        with TemporaryDirectory() as directory:
-            root = historical_tree(A135, directory)
-            for name, expected in pins["sources"].items():
-                self.assertEqual(hashlib.sha256((root / name).read_bytes()).hexdigest(), expected)
-            self.assertEqual(pins["sources"][HEADER], pins["candidate_header_sha256"])
-            self.assertNotEqual(pins["sources"][HEADER], pins["baseline_header_sha256"])
-            with self.assertRaisesRegex(RuntimeError, "Source pin mismatch"):
-                source_guard(root, "baseline", "a135252")
+        tracked = subprocess.check_output(["git", "-C", str(ROOT), "ls-files", "src"], text=True)
+        self.assertEqual(set(pins["sources"]), set(tracked.splitlines()))
+        for name, expected in pins["sources"].items():
+            self.assertEqual(hashlib.sha256((ROOT / name).read_bytes()).hexdigest(), expected)
+        self.assertEqual(pins["sources"][HEADER], pins["candidate_header_sha256"])
+        self.assertNotEqual(pins["sources"][HEADER], pins["baseline_header_sha256"])
+        with self.assertRaisesRegex(RuntimeError, "Source pin mismatch"):
+            source_guard(ROOT, "baseline", "a135252")

    def test_named_baseline_accepts_only_reversed_header(self):
        pins = json.loads((HERE / "data/les_flow/source_pins_a135252.json").read_text())
        with TemporaryDirectory() as directory:
-            origin = historical_tree(A135, Path(directory) / "origin")
-            root = Path(directory) / "fixture"
+            root = Path(directory)
            for name in pins["sources"]:
                dest = root / name
                dest.parent.mkdir(parents=True, exist_ok=True)
-                shutil.copyfile(origin / name, dest)
+                shutil.copyfile(ROOT / name, dest)
            subprocess.run(["git", "apply", "--reverse", str(HERE / "les/data/smagorinsky_normalization.patch")], cwd=root, check=True)
            with patch("run_uniform_les_flow.subprocess.check_output", return_value="baseline-fixture\n"):
                source_guard(root, "baseline", "a135252")
@@ -135,6 +135,6 @@ def test_d2_entrypoint_compiles_in_isolation(tmp_path, model, shift, storage):
            source=re.sub(r"(#define\s+"+key+r"\s+)\S+",lambda match: match[1]+str(value),source)
        path.write_text(source)
    result=subprocess.run([nvcc,"-ptx","-arch=sm_70","-std=c++17",
-                           str(copy/"uniform.cu"),"-o",str(tmp_path/"kernel.ptx")],
+                           str(copy/"kernel_v2.cu"),"-o",str(tmp_path/"kernel.ptx")],
                          capture_output=True,text=True)
    assert result.returncode == 0, result.stderr
@@ -9,7 +9,7 @@ import pytest

 import CelerisLab
 from CelerisLab.config import LBMConfig, load_lbm_config
-from CelerisLab.cuda import compiler
+from CelerisLab.cuda import compiler_v2 as compiler
 from CelerisLab.common.checkpoint import save_checkpoint, load_checkpoint
 from .test_numerical_operators import compile_run
 from .test_state_regressions import device, make_field, manager
@@ -18,7 +18,7 @@ from .test_state_regressions import device, make_field, manager
 def test_planar_helper_moments_and_known_populations(tmp_path):
    root = Path(__file__).resolve().parents[2]
    assert Path(CelerisLab.__file__).resolve() == root/'src/CelerisLab/__init__.py'
-    assert Path(compiler.kernel_path('uniform.cu')).resolve() == root/'src/CelerisLab/lbm/kernels/uniform.cu'
+    assert Path(compiler.kernel_path('kernel_v2.cu')).resolve() == root/'src/CelerisLab/lbm/kernels/kernel_v2.cu'
    lines = compile_run(tmp_path, 'planar_velocity.cpp', dict(WALL_NORMAL_VELOCITY='0.0f'))
    cx = np.array([0,1,-1,0,0,1,-1,1,-1])
    cy = np.array([0,0,0,1,-1,1,-1,-1,1])
@@ -162,7 +162,7 @@ class ProjectedKernelTests(unittest.TestCase):
        targets = child.owner.ravel()[p.ghost_fine]
        donors = parent.owner.ravel()[p.ghost_coarse]
        source = (Path(__file__).resolve().parents[2] /
-                  "src/CelerisLab/lbm/kernels/step/block_transfer.cu").read_text()
+                  "src/CelerisLab/lbm/kernels/blocks.cu").read_text()
        kernel = source.split('extern "C" __global__ void BlockExplodeProjected(',1)[1]
        kernel = 'void BlockExplodeProjected(' + kernel.split('// Targets are receiver-local SOURCE slots',1)[0]
        def array(name, values, kind):
@@ -29,7 +29,7 @@ def backend(**options):
            name=name, side_effect=lambda *a, **kw: calls.append((name, a, kw)))
        return module

-    host.compiler.build_module.side_effect = lambda *a, **k: load(None)
+    host.compiler.load_module.side_effect = load
    host.cuda.memcpy_htod.side_effect = lambda ptr, a: storage.__setitem__(id(ptr), np.array(a, copy=True))
    host.cuda.memcpy_dtoh.side_effect = lambda a, ptr: np.copyto(a, storage[id(ptr)])
    cfg = options.pop('cfg', CFG)
@@ -146,6 +146,7 @@ class ProjectedPairDispatch(unittest.TestCase):
        for cfg, changes in [(CFG, {'explosion':'unknown'}),
                             (CFG, {'coarse_boundary':'slot_reservoir'}),
                             (CFG, {'coarse_boundary':'channel'}),
+                             (CFG, {'level_blocks':[None,None], 'refinements':[(6,6,26,26)]}),
                             (CFG, {'level_objects':[(),(object(),)]}),
                             (CFG, {'level_jets':[(),(object(),)]}),
                             (replace(CFG,collision='MRT'), {}),
@@ -154,8 +155,9 @@ class ProjectedPairDispatch(unittest.TestCase):
                host.FixedPatchStepper(cfg, **dict(options, **changes))
        context.activate.assert_not_called()
        host.cuda.mem_alloc.assert_not_called()
-        # The constant path still allows MRT.
-        for options in ({'cfg':replace(CFG,collision='MRT')},):
+        # The legacy constant path still allows MRT and dense storage.
+        for options in ({'cfg':replace(CFG,collision='MRT')},
+                        {'level_blocks':[None,None], 'refinements':[(6,6,26,26)]}):
            run, _, _, _, _ = backend(**options)
            run.close()

@@ -25,7 +25,7 @@ def backend(observe=True, obj=None):
    module = Mock()
    module.get_global.return_value = (Mock(), 20)
    module.get_function.side_effect = lambda name: Mock(side_effect=lambda *a, **k: calls.append((name, a, k)))
-    host.compiler.build_module.return_value = module
+    host.compiler.load_module.return_value = module
    host.cuda.mem_alloc.side_effect = lambda n: Mock(size=n)
    host.cuda.memcpy_dtoh.side_effect = lambda array, pointer: array.fill(0)
    ctx = Mock(activate=nullcontext, sm_arch='sm_70')
@@ -85,7 +85,7 @@ class RawExchange(unittest.TestCase):

    def test_four_level_schedule_publication_and_fixed_storage(self):
        run, host, calls, arrays = backend()
-        b = run._device_boundaries[3]
+        b = run._level_boundaries[3]
        rows = len(b.tables['cell'])
        self.assertEqual(b.last_raw.size, rows * 16)
        self.assertEqual(b.interval_integral.size, rows * 32)
@@ -128,7 +128,7 @@ class RawExchange(unittest.TestCase):

    def test_disabled_has_nulls_and_no_observation_allocation_or_launch(self):
        run, host, calls, _ = backend(False)
-        b = run._device_boundaries[3]
+        b = run._level_boundaries[3]
        self.assertIs(b.last_raw, b.interval_integral)
        self.assertIs(b.interval_integral, None)
        run.step()
@@ -153,6 +153,14 @@ class RawExchange(unittest.TestCase):
        run._stream.synchronize.side_effect = None
        run.close()

+    def test_dense_observation_rejected_before_device_access(self):
+        host = host_module()
+        ctx = Mock()
+        with self.assertRaisesRegex(ValueError, 'blocked body'):
+            host.FixedPatchStepper(LBMConfig(nx=32, ny=32, viscosity=.1), (4, 4, 28, 28),
+                                   context=ctx, fine_objects=[body(radius=2.15)], observe_raw_exchange=True)
+        ctx.activate.assert_not_called()
+
    def test_all_depth_weights(self):
        for level in (1, 2, 3, 7):
            self._check_weights_split_and_window(level)
@@ -5,7 +5,6 @@ import pytest

 ROOT = Path(__file__).resolve().parents[2]
 sys.path.insert(0, str(ROOT / "tests/validation"))
-sys.path.insert(0, str(ROOT / "tests/unit"))
 import re100_helpers as h
 import periodic as p

@@ -63,16 +62,13 @@ def test_periodic_synthetic():
    assert not p.assess(rows()[:100])["periodic"]


-def test_production_unmodified(tmp_path):
-    from historical_tree import historical_tree
-
+def test_production_unmodified():
    m = json.loads(
        (ROOT / "tests/validation/data/dfg_re100/source_manifest.json").read_text()
    )
    assert m["commit"] == h.PIN
-    origin = historical_tree(m["commit"], tmp_path)
    for name, digest in m["production_sha256"].items():
-        assert hashlib.sha256((origin / name).read_bytes()).hexdigest() == digest, name
+        assert hashlib.sha256((ROOT / name).read_bytes()).hexdigest() == digest, name


 def test_reference():
@@ -7,10 +7,7 @@ import numpy as np
 import unittest

 sys.path.insert(0, str(Path(__file__).resolve().parents[1] / "validation"))
-sys.path.insert(0, str(Path(__file__).resolve().parent))
 import run_refined_dfg as r
-from historical_tree import historical_tree
-from tempfile import TemporaryDirectory

 h, periodic = r.helpers(r.DEFAULT_SOURCE)

@@ -20,24 +17,11 @@ def test_source_pin():
    assert historical["commit"].startswith("cd3a657")
    mrt_history = json.loads((r.HERE / "data/refined_dfg/source_manifest_mrt.json").read_text())
    assert mrt_history["commit"].startswith("3ea131e")
-    with unittest.TestCase().assertRaisesRegex(RuntimeError, "Source mismatch"):
-        r.verify_source(r.DEFAULT_SOURCE, "MRT", "window")
+    assert r.verify_source(r.DEFAULT_SOURCE, "MRT", "window")["commit"].startswith("a233c90")
    with unittest.TestCase().assertRaisesRegex(RuntimeError, "Source mismatch"):
        r.verify_source(r.DEFAULT_SOURCE, "MRT")
    with unittest.TestCase().assertRaisesRegex(RuntimeError, "Source mismatch"):
        r.verify_source(r.DEFAULT_SOURCE, "SRT")
-    with TemporaryDirectory() as directory:
-        for collision, mode, prefix in (
-            ("SRT", "last_base_step", "cd3a657"),
-            ("MRT", "last_base_step", "3ea131e"),
-            ("MRT", "window", "a233c90"),
-        ):
-            name = {"SRT": "source_manifest.json", "MRT": "source_manifest_mrt.json"}[collision]
-            name = "source_manifest_window.json" if mode == "window" else name
-            commit = json.loads((r.HERE / "data/refined_dfg" / name).read_text())["commit"]
-            assert commit.startswith(prefix)
-            root = historical_tree(commit, Path(directory) / commit)
-            assert r.verify_source(root, collision, mode)["commit"] == commit


 def test_matched_units_and_planes():
@@ -157,32 +141,16 @@ def test_axis_specific_sampling_phase():
    np.testing.assert_allclose(fine_y - ((k + 43) - 0.5) * 0.00125, 0, atol=1e-15)


-def frozen_manifest(observation_mode):
-    """Frozen MRT manifest JSON as recorded, without hashing the working tree."""
-    name = "source_manifest_window.json" if observation_mode == "window" else "source_manifest_mrt.json"
-    return json.loads((r.HERE / "data/refined_dfg" / name).read_text())
-
-
-def selected_manifest(collision, observation_mode):
-    """The frozen manifest verify_source selects for (collision, observation_mode).
-
-    Source identity is owned by test_source_pin; this reads no checkout files.
-    """
-    return r.source_manifest(collision, observation_mode)
-
-
 def test_full_requires_completed_pilot():
    with unittest.TestCase().assertRaisesRegex(ValueError, "pilot-results"):
-        r.verify_pilot(None, frozen_manifest("window"))
+        r.verify_pilot(None, r.verify_source(r.DEFAULT_SOURCE, "MRT", "window"))


 def pilot_fixture(path, observation_mode="window", jet_scenario="none", case_selection="both", uniform_diameter=80, placement="wide"):
    """Synthetic contract fixture only; never a physical result or run receipt."""
-    # Receipt-logic fixture: the frozen manifest verify_source returns for a matching tree.
-    # Source identity itself is checked only by test_source_pin.
    manifest = (r.verify_source(r.DEFAULT_SOURCE, "MRT", "window", jet_scenario)
-                if observation_mode == "window" and jet_scenario != "none" else
-                frozen_manifest(observation_mode))
+                if observation_mode == "window" else
+                json.loads((r.HERE / "data/refined_dfg/source_manifest_mrt.json").read_text()))
    selection = dict(case_selection=case_selection, uniform_diameter=uniform_diameter, placement=placement)
    (path / "plan.json").write_text(
        json.dumps(
@@ -380,7 +348,7 @@ def test_record_actual_checkout_separately():
    actual = checkout_provenance(r.DEFAULT_SOURCE)
    assert len(actual["checkout_revision"]) == 40
    assert isinstance(actual["checkout_dirty"], bool)
-    assert selected_manifest("MRT", "window")["production_origin_commit"].startswith("d930")
+    assert r.verify_source(r.DEFAULT_SOURCE, "MRT", "window")["production_origin_commit"].startswith("d930")


 def test_window_plan_preserves_physics():
@@ -390,7 +358,7 @@ def test_window_plan_preserves_physics():
        assert baseline.pop("observation_mode") == "last_base_step"
        assert window.pop("observation_mode") == "window"
        assert baseline == window
-        assert selected_manifest(collision, "window")["observation_mode"] == "window"
+        assert r.verify_source(r.DEFAULT_SOURCE, collision, "window")["observation_mode"] == "window"
    with unittest.TestCase().assertRaises(ValueError):
        r.verify_source(r.DEFAULT_SOURCE, "MRT", "unknown")

@@ -22,29 +22,23 @@ class RefinementMRTHost(unittest.TestCase):

    def test_srt_statements_and_guards_preserved(self):
        """MRT is an additive compile-time branch; SRT statements and other guards remain."""
-        blocks = (KERNELS / "blocked.cu").read_text() + (KERNELS / "step/block_step.cu").read_text()
+        blocks = (KERNELS / "blocks.cu").read_text()
        channel = (KERNELS / "step/block_channel.cu").read_text()
-        block_collision = (KERNELS / "operators/block_collision.cuh").read_text()
-        self.assertIn('#include "operators/block_collision.cuh"', blocks)
-        self.assertEqual(blocks.count("block_collide(f);"), 1)
+        refinement = (KERNELS / "refinement.cu").read_text()
        for text in ("compute_feq_from_momentum(drho, jx, jy, ux, uy, feq);",
                     "compute_feq(rho, ux, uy, feq);",
                     "collide_srt_no_force(f, feq, d_params.omega);"):
-            self.assertIn(text, block_collision)
+            self.assertIn(text, blocks)
        self.assertIn("collide_srt(f, feq, forcing, omega);", channel)
-        for source in (blocks,):
+        self.assertIn("collide_srt_no_force(f, feq, d_params.omega);", refinement)
+        for source in (blocks, refinement):
            self.assertIn("(COLLISION_MODEL != 0 && COLLISION_MODEL != 2)", source)
            self.assertIn("USE_LES", source)
            self.assertIn("STORE_PRECISION != 0", source)
            self.assertIn('#if COLLISION_MODEL == 2\n#include "operators/collision_mrt.cuh"\n#endif', source)
-        # One shared unforced MRT body; each blocked/patch path calls it exactly once.
-        mrt = (KERNELS / "operators/collision_mrt.cuh").read_text()
-        self.assertEqual(mrt.count("void collide_mrt_unforced("), 1)
-        for source in (blocks + "\n" + block_collision, channel):
-            self.assertEqual(source.count("collide_mrt_unforced(f);"), 1)
-            self.assertNotIn("collide_mrt_no_force(", source)
-            self.assertNotIn("compute_rho_u(f, rho, ux, uy);\n            collide_mrt", source)
-            self.assertNotRegex(source, r"compute_rho_u\(f, rho, ux, uy\);\s*collide_mrt")
+        self.assertIn("USE_DDF_SHIFTING ||", refinement)  # dense/mixed shifted stays rejected
+        for source in (blocks, channel, refinement):
+            self.assertEqual(source.count("collide_mrt_no_force("), 1)


 if __name__ == "__main__": unittest.main()
@@ -22,13 +22,12 @@ def host_module():
    tree.body=[node for node in tree.body if not (
        isinstance(node,ast.Import) and any(a.name=='pycuda.driver' for a in node.names)
        or isinstance(node,ast.ImportFrom) and node.module in (
-            'cuda','curved_links'))]
+            'cuda','block_stepper','curved_links'))]
    module=types.ModuleType('CelerisLab.lbm._test_host_refinement')
    module.__package__='CelerisLab.lbm'
    from CelerisLab.body.jets import surface_link_velocity
-    module.__dict__.update(cuda=Mock(),compiler=Mock(),
+    module.__dict__.update(cuda=Mock(),compiler=Mock(),MAX_GRID_DIM=65535,
                           surface_link_velocity=surface_link_velocity)
-    module.cuda.mem_get_info.return_value=(1<<40,1<<40)  # free-memory preflight passes
    exec(compile(tree,str(path),'exec'),module.__dict__)
    return module

@@ -141,7 +140,7 @@ class Gate(unittest.TestCase):
    def test_sparse_parent_tables_preserve_inherited_kinds(self):
        child,kind=r.nest_patch_set(CFG,self.parent,CHILDREN)
        layout=self.parent.fine_layout((4,4));order=np.argsort(child.coalesce_coarse)
-        donors,targets,bkind=host.blocked_coarse_tables(child,layout,order,composed_kind=kind)
+        donors,targets,bkind=host.blocked_base_tables(child,layout,order,composed_kind=kind)
        self.assertTrue((donors < layout.hole).all())
        self.assertTrue((targets < np.prod(layout.blocked_shape)).all())
        np.testing.assert_array_equal(bkind[layout.owner[layout.owned]],kind[layout.owned])
@@ -158,10 +157,10 @@ class Gate(unittest.TestCase):
        hole=int(np.flatnonzero(~layout.owned.ravel())[0])
        bad=child.ghost_coarse.copy();bad[0]=hole
        with self.assertRaisesRegex(ValueError,'owned uncovered'):
-            host.blocked_coarse_tables(replace(child,ghost_coarse=bad),layout,order,composed_kind=kind)
+            host.blocked_base_tables(replace(child,ghost_coarse=bad),layout,order,composed_kind=kind)
        bad=child.coalesce_coarse.copy();bad[0]=hole
        with self.assertRaisesRegex(ValueError,'owned uncovered'):
-            host.blocked_coarse_tables(replace(child,coalesce_coarse=bad),layout,order,composed_kind=kind)
+            host.blocked_base_tables(replace(child,coalesce_coarse=bad),layout,order,composed_kind=kind)


 def bodies_and_jets(speed=.002):
@@ -174,10 +173,8 @@ def bodies_and_jets(speed=.002):
    return bodies,jets


-def levels(parents=PARENTS,children=CHILDREN,blocks=((4,4),)*3,objects=((),(),()),jets=((),(),())):
-    """Level-list constructor keywords for the L0 set, L1 set and L2 set used here."""
-    return dict(refinements=[parents,children],level_blocks=list(blocks),
-                level_objects=list(objects),level_jets=list(jets))
+KW=dict(base_block=(4,4),fine_block=(4,4),child_block=(4,4),
+        child_rectangles_in_parent=CHILDREN)


 def mock_backend(*, children=CHILDREN, bodies=False, fail_compile=False, fail_upload=False):
@@ -194,9 +191,9 @@ def mock_backend(*, children=CHILDREN, bodies=False, fail_compile=False, fail_up
        gpu.get_function.side_effect=lambda name: Mock(name=name)
        gpu.get_global.side_effect=lambda name: (Mock(),20 if name=='d_params' else 16)
        return gpu
-    module.compiler.build_module.side_effect=lambda *a,**k: load(None)
+    module.compiler.load_module.side_effect=load
    if fail_compile:
-        module.compiler.build_module.side_effect=[load(None),load(None),RuntimeError('L2 compile failed')]
+        module.compiler.compile_kernel.side_effect=['first','second',RuntimeError('L2 compile failed')]
    if fail_upload:
        def upload(*args):
            if len(allocations)>20:
@@ -204,8 +201,9 @@ def mock_backend(*, children=CHILDREN, bodies=False, fail_compile=False, fail_up
        module.cuda.memcpy_htod.side_effect=upload
    objects,jets=bodies_and_jets() if bodies else ((),())
    try:
-        run=module.FixedPatchStepper(CFG,context=context,
-            **levels(children=children,objects=((),(),objects),jets=((),(),jets)))
+        run=module.FixedPatchStepper(CFG,PARENTS,context=context,
+            child_objects=objects,child_jets=jets,
+            **dict(KW,child_rectangles_in_parent=children))
    except RuntimeError:
        for allocation in allocations:
            allocation.free.assert_called_once_with()
@@ -216,29 +214,28 @@ def mock_backend(*, children=CHILDREN, bodies=False, fail_compile=False, fail_up
 class ConstructorAndSchedule(unittest.TestCase):
    def test_invalid_inputs_fail_before_context_or_allocations(self):
        context=Mock()
-        b=(4,4)
-        cases=[dict(children=(8,8,16,16)),dict(children=[]),
-               dict(children=[(16,16,32)]),
-               dict(children=[(16.5,16,32,40)]),
-               dict(children=[(16,16,32,24),(16,27,32,55)]),
-               dict(children=[(40,16,48,40)]),
-               dict(children=[(8,16,32,40)]),
-               dict(blocks=(b,b,None)),dict(blocks=(b,None,b)),dict(blocks=(None,b,b)),
-               dict(blocks=(b,b,(5,4))),dict(blocks=(b,(5,4),b)),
-               dict(blocks=(b,b,(4,))),dict(children=(16,16,32,40))]
+        cases=[dict(child=(8,8,16,16)),dict(child_rectangles_in_parent=[]),
+               dict(child_rectangles_in_parent=[(16,16,32)]),
+               dict(child_rectangles_in_parent=[(16.5,16,32,40)]),
+               dict(child_rectangles_in_parent=[(16,16,32,24),(16,27,32,55)]),
+               dict(child_rectangles_in_parent=[(40,16,48,40)]),
+               dict(child_rectangles_in_parent=[(8,16,32,40)]),
+               dict(child_block=None),dict(fine_block=None),dict(base_block=None),
+               dict(child_block=(5,4)),dict(fine_block=(5,4)),
+               dict(child_block=(4,)),dict(child_rectangles_in_parent=(16,16,32,40))]
        for changes in cases:
            with self.subTest(changes=changes),self.assertRaises((ValueError,TypeError)):
-                host.FixedPatchStepper(CFG,context=context,**levels(**changes))
+                host.FixedPatchStepper(CFG,PARENTS,context=context,**dict(KW,**changes))
        with self.assertRaisesRegex(ValueError,'margin'):
-            host.FixedPatchStepper(CFG,context=context,**levels(parents=[(1,4,20,28),(23,4,44,28)]))
+            host.FixedPatchStepper(CFG,[(1,4,20,28),(23,4,44,28)],context=context,**KW)
        objects,jets=bodies_and_jets()
-        with self.assertRaisesRegex(ValueError,'only on the finest level'):
-            host.FixedPatchStepper(CFG,context=context,**levels(objects=((),objects,())))
+        with self.assertRaisesRegex(ValueError,'only on L2'):
+            host.FixedPatchStepper(CFG,PARENTS,context=context,fine_objects=objects,**KW)
        from CelerisLab.body.jets import Jet,Tube,Parallel
        spanning=Jet(Tube((10.5,14.),0.,25.,.3),Parallel(0.),amplitude=0.)
        with self.assertRaisesRegex(ValueError,'one jet'):
-            host.FixedPatchStepper(CFG,context=context,
-                                  **levels(objects=((),(),objects),jets=((),(),[spanning])))
+            host.FixedPatchStepper(CFG,PARENTS,context=context,child_objects=objects,
+                                  child_jets=[spanning],**KW)
        context.activate.assert_not_called()
        host.cuda.mem_alloc.assert_not_called()

@@ -316,23 +313,21 @@ class ConstructorAndSchedule(unittest.TestCase):



-class RemovedLegacyAPI(unittest.TestCase):
-    """The pre-level-list constructor/initialize forms are gone and fail loudly."""
+class InitialLevels(unittest.TestCase):
+    """initialize() entry adapter: positional per-level arrays or the legacy names, never both."""

-    def test_removed_constructor_forms_raise_type_error(self):
-        import inspect
-        context = Mock()
-        with self.assertRaises(TypeError):  # positional bounds
-            host.FixedPatchStepper(CFG, PARENTS, context=context)
-        removed = dict(bounds=PARENTS, fine_block=(4, 4), base_block=(4, 4), fine_objects=(),
-                       jets=(), child=(8, 8, 16, 16), child_block=(4, 4), child_objects=(),
-                       child_jets=(), child_rectangles_in_parent=CHILDREN)
-        for name, value in removed.items():
-            with self.subTest(keyword=name), self.assertRaises(TypeError):
-                host.FixedPatchStepper(CFG, context=context, **levels(), **{name: value})
-        context.activate.assert_not_called()
-        self.assertEqual(list(inspect.signature(host.FixedPatchStepper.initialize).parameters),
-                         ["self", "level_g"])  # no coarse_g/fine_g/child_g
+    def test_forms(self):
+        a, b, c = object(), object(), object()
+        f = host._initial_levels
+        self.assertEqual(f((a, b), None, None, None), (a, b))
+        self.assertEqual(f((a, b, c), None, None, None), (a, b, c))
+        self.assertEqual(f((a, b), None, None, c), (a, b, c))      # old initialize(c, f, child_g=g)
+        self.assertEqual(f((), a, b, None), (a, b))
+        self.assertEqual(f((), a, b, c), (a, b, c))
+        for args in (((a,), None, None, c), ((a, b, c), None, None, c), ((a,), a, b, None),
+                     ((), a, None, None), ((), None, b, c)):
+            with self.subTest(args=args), self.assertRaises(TypeError):
+                f(*args)


 if __name__=='__main__':
@@ -86,7 +86,7 @@ class RuntimeChildRemap(unittest.TestCase):
            before = run.download()
            old_allocations = list(allocations)
            if failure == 'construction':
-                operation = patch.object(module.compiler, 'build_module', side_effect=RuntimeError('candidate compile'))
+                operation = patch.object(module.compiler, 'compile_kernel', side_effect=RuntimeError('candidate compile'))
            else:
                initialize = type(run).initialize
                def fail(candidate, *arrays):
@@ -54,20 +54,21 @@ class ShiftedJets(unittest.TestCase):
                (Jet(Tube((10.,14.),0.,26.,.3),Parallel(0.)),))]
        for objects,jets in cases:
            with self.subTest(objects=len(objects),jets=len(jets)),self.assertRaises(ValueError):
-                module.FixedPatchStepper(CFG,context=None,refinements=[backend.PARENTS,A],
-                    level_blocks=[(4,4)]*3,level_objects=[(),(),objects],level_jets=[(),(),jets])
+                module.FixedPatchStepper(CFG,backend.PARENTS,context=None,
+                    base_block=(4,4),fine_block=(4,4),child_block=(4,4),
+                    child_rectangles_in_parent=A,child_objects=objects,child_jets=jets)
        for blocks in ((None,None,None),((4,4),None,(4,4))):
            with self.assertRaises(ValueError):
-                module.FixedPatchStepper(CFG,context=None,refinements=[backend.PARENTS,A],
-                    level_blocks=list(blocks),level_objects=[(),(),(body(),)],
-                    level_jets=[(),(),(jet(),)])
+                module.FixedPatchStepper(CFG,backend.PARENTS,context=None,
+                    base_block=blocks[0],fine_block=blocks[1],child_block=blocks[2],
+                    child_rectangles_in_parent=A,child_objects=(body(),),child_jets=(jet(),))
        module.cuda.mem_alloc.assert_not_called()
-        module.compiler.build_module.assert_not_called()
+        module.compiler.compile_kernel.assert_not_called()

    def test_switch_upload_only_then_raw_remap_keeps_live_velocity(self):
        run,module,allocations,storage,objects=ready();clone=None
        try:
-            boundary=run._device_boundaries[2]
+            boundary=run._level_boundaries[2]
            before=run._download_raw()
            original=boundary.tables['surface_velocity'].copy()
            module.cuda.memcpy_htod.reset_mock()
@@ -79,7 +80,7 @@ class ShiftedJets(unittest.TestCase):
            for a,b in zip(before.values,run._download_raw().values):self.assertEqual(a.tobytes(),b.tobytes())
            clone=run.remap_child(B,child_objects=objects)
            self.assertEqual(clone._level_jets[2][0].amplitude,-jet().amplitude)
-            np.testing.assert_array_equal(clone._device_boundaries[2].tables['surface_velocity'],
+            np.testing.assert_array_equal(clone._level_boundaries[2].tables['surface_velocity'],
                                          boundary.tables['surface_velocity'])
            self.assertEqual(clone.step_count,7)
            saved=run._download_raw()
@@ -92,12 +93,12 @@ class ShiftedJets(unittest.TestCase):
    def test_invalid_amplitude_does_not_upload_or_commit(self):
        run,module,_,_,_=ready()
        try:
-            before=run._device_boundaries[2].tables['surface_velocity'].copy()
+            before=run._level_boundaries[2].tables['surface_velocity'].copy()
            module.cuda.memcpy_htod.reset_mock()
            for amplitude in (float('nan'),float('inf'),1.):
                with self.assertRaises(ValueError):run.set_jet_amplitude(0,amplitude,level=2)
            module.cuda.memcpy_htod.assert_not_called()
-            np.testing.assert_array_equal(before,run._device_boundaries[2].tables['surface_velocity'])
+            np.testing.assert_array_equal(before,run._level_boundaries[2].tables['surface_velocity'])
        finally:run.close()


@@ -1,4 +1,7 @@
 """CPU host contracts; mocked CUDA executes real lifecycle, not numerical kernels."""
+import ast
+import types
+from pathlib import Path
 from contextlib import nullcontext
 from dataclasses import replace
 import unittest
@@ -75,15 +78,13 @@ class ShiftedStorage(unittest.TestCase):

    def test_shifted_density_and_tag_validation(self):
        old,new = topology(A),topology(B)
-        a = tuple(np.where(m[None], np.float32(-.001), np.float32(np.nan)).repeat(9,axis=0)
-                  for m in old.leaf_masks)  # FP32 under NumPy 1 and 2 promotion rules
+        a = tuple(np.where(m[None], np.float32(-.001), np.nan).repeat(9,axis=0) for m in old.leaf_masks)
        remap(old,new,a)
        with self.assertRaisesRegex(ValueError,'representation'): remap(old,new,a,False)
        with self.assertRaisesRegex(ValueError,'FP32'): remap(old,new,tuple(x.astype(np.float64) for x in a))
        a[0][:,old.leaf_masks[0]] = -1
        with self.assertRaisesRegex(ValueError,'positive density'): remap(old,new,a)
-        rest = tuple(np.where(m[None],np.float32(0),np.float32(np.nan)).repeat(9,axis=0)
-                     for m in old.leaf_masks)
+        rest = tuple(np.where(m[None],np.float32(0),np.nan).repeat(9,axis=0) for m in old.leaf_masks)
        remap(old,new,rest)
        with self.assertRaisesRegex(ValueError,'positive density'):
            remap_synchronized(old,new,rest,base_count=7,level_steps=(7,14,28),phase='postcollision_g')
@@ -125,24 +126,33 @@ class ShiftedStorage(unittest.TestCase):
            with self.assertRaisesRegex(ValueError,'mixed'): run._download_raw()
        finally: run.close()

-    def test_missing_tiles_rejected_before_compilation(self):
+    def test_dense_or_mixed_shifted_rejected_before_compilation(self):
        module=backend.host_module()
        cfg=replace(backend.CFG,ddf_shifting=True)
-        for blocks in ([None,None],[(4,4),None],[None,(4,4)]):
-            with self.assertRaisesRegex(ValueError,'must be a tile'):
-                module.FixedPatchStepper(cfg,context=None,refinements=[(4,4,20,28)],
-                                         level_blocks=blocks)
-        module.compiler.build_module.assert_not_called()
+        for kwargs in ({},dict(base_block=(4,4)),dict(fine_block=(4,4))):
+            with self.assertRaisesRegex(ValueError,'all-blocked|base_block requires fine_block'):
+                module.FixedPatchStepper(cfg,(4,4,20,28),context=None,**kwargs)
+        module.compiler.compile_kernel.assert_not_called()

    def test_uniform_block_codec_and_raw_snapshot(self):
-        module = backend.host_module()  # single blocked level: FixedPatchStepper(refinements=[])
-        cuda, compiler = module.cuda, module.compiler
+        path = Path(backend.r.__file__).with_name('block_stepper.py')
+        tree = ast.parse(path.read_text())
+        tree.body = [n for n in tree.body if not (
+            isinstance(n, ast.Import) and any(a.name == 'pycuda.driver' for a in n.names)
+            or isinstance(n, ast.ImportFrom) and n.module == 'cuda')]
+        module = types.ModuleType('CelerisLab.lbm._test_block_host')
+        module.__package__ = 'CelerisLab.lbm'
+        cuda, compiler = Mock(), Mock()
+        module.__dict__.update(cuda=cuda,compiler=compiler)
+        exec(compile(tree,str(path),'exec'),module.__dict__)
+        cuda.mem_get_info.return_value=(10**9,10**9)
        cuda.mem_alloc.side_effect=lambda size: Mock()
        gpu=Mock();gpu.get_global.return_value=(Mock(),20)
-        compiler.build_module.return_value=gpu
+        compiler.load_module.return_value=gpu
        context=Mock(activate=nullcontext,sm_arch='sm_70')
+        layout=module.BlockLayout2D.build(16,16,4,4)
        cfg=replace(backend.CFG,nx=16,ny=16,ddf_shifting=True)
-        run=module.FixedPatchStepper(cfg,context=context,refinements=[],level_blocks=[(4,4)])
+        run=module.PeriodicBlockStepper(cfg,layout,context=context)
        storage={}
        cuda.memcpy_htod.side_effect=lambda p,a: storage.__setitem__(id(p),np.array(a,copy=True))
        cuda.memcpy_dtoh.side_effect=lambda a,p: np.copyto(a,storage[id(p)])
@@ -153,8 +163,7 @@ class ShiftedStorage(unittest.TestCase):
            self.assertTrue(raw.shifted)
            self.assertEqual(len(raw.values),1)
            self.assertTrue((raw.values[0][1]==np.float32(2**-35)).all())
-            (down,)=run.download()
-            np.testing.assert_array_equal(down,physical)
+            np.testing.assert_array_equal(run.download(),physical)
        finally: run.close()

    def test_passive_shifted_body_raw_remap_preserves_support(self):
@@ -36,10 +36,12 @@ def make_simulation(monkeypatch):
    sim = module.Simulation.__new__(module.Simulation)
    sim.ctx = Owner()
    sim._closed = False
+    sim._build_dir = "private-build"
    for name in ("stepper", "bodies", "field", "_module", "_stream"):
        setattr(sim, name, object())
    monkeypatch.setattr(module, "cuda", SimpleNamespace(
        Context=SimpleNamespace(synchronize=lambda: None)))
+    monkeypatch.setattr(module.compiler, "cleanup_build_dir", lambda path: None)
    return sim


@@ -55,7 +57,7 @@ def test_close_inside_activation_preserves_resources(monkeypatch):
    assert sim._closed and sim.ctx.is_closed


-@pytest.mark.parametrize("failure", ["synchronize", "detach"])
+@pytest.mark.parametrize("failure", ["synchronize", "detach", "directory"])
 def test_close_failure_is_retryable(monkeypatch, failure):
    sim = make_simulation(monkeypatch)
    if failure == "detach":
@@ -66,12 +68,17 @@ def test_close_failure_is_retryable(monkeypatch, failure):
            calls.append(args)
            if len(calls) == 1:
                raise RuntimeError(failure + " failed")
-        monkeypatch.setattr(module.cuda.Context, "synchronize", fail_once)
+        if failure == "synchronize":
+            monkeypatch.setattr(module.cuda.Context, "synchronize", fail_once)
+        else:
+            monkeypatch.setattr(module.compiler, "cleanup_build_dir", fail_once)
    with pytest.raises(RuntimeError, match="failed"):
        sim.close()
    assert not sim._closed
+    assert sim._build_dir == "private-build"
    sim.close()
    assert sim._closed and sim.ctx.is_closed
+    assert sim._build_dir is None
    sim.close()


@@ -1,218 +0,0 @@
-"""Host contract of FixedPatchStepper(refinements=[]): one blocked level (mocked driver).
-
-Per step: halo refresh -> static BFL (circles only) -> BlockStep with a null
-kind, one stream wait per step(n) call, NaN spare after initialize. An upload,
-launch, wait or download failure fails the instance and keeps its step count;
-close retains resources for retry when context activation fails.
-"""
-from contextlib import nullcontext
-from unittest.mock import MagicMock, Mock
-import unittest
-
-import numpy as np
-
-from CelerisLab.body.geometry.circle import CircleGeometry
-from CelerisLab.body.objects import SimObject
-from CelerisLab.config import LBMConfig
-from tests.unit.test_refinement_three_level_multipatch import host_module
-
-
-def circle(center=(15.5, 11.5), radius=4.3):
-    return SimObject(0, geometry=CircleGeometry(*center, radius), center=center, radius=radius)
-
-
-def backend(objects=(), failing_kernel=None, initialize=True):
-    host, log = host_module(), []
-    host.armed = False  # set True to make failing_kernel raise
-
-    def launch(name):
-        def run(*a, **k):
-            log.append((name, a))
-            if name == failing_kernel and host.armed:
-                raise RuntimeError(name + " launch failed")
-        return run
-
-    module = Mock()
-    module.get_global.return_value = (Mock(), 20)
-    module.get_function.side_effect = lambda name: Mock(side_effect=launch(name))
-    host.compiler.build_module.return_value = module
-    host.cuda.mem_alloc.side_effect = lambda size: Mock(size=size)
-    host.cuda.memcpy_htod.side_effect = lambda pointer, array: log.append(("upload", (pointer, array)))
-    run = host.FixedPatchStepper(LBMConfig(nx=32, ny=24, viscosity=.1),
-                                 context=Mock(activate=nullcontext, sm_arch="sm_70"),
-                                 refinements=[], level_blocks=[(12, 16)], level_objects=[objects])
-    run._stream.synchronize.side_effect = lambda: log.append(("sync", ()))
-    log.clear()
-    if initialize:
-        run.initialize(np.full(run._levels[0].shape, 1 / 9, np.float32))
-    return run, log, host
-
-
-def state():
-    return np.full((9, 24, 32), 1 / 9, np.float32)
-
-
-class SingleLevel(unittest.TestCase):
-    def test_spare_starts_nan_and_current_holds_state(self):
-        run, log, _ = backend()
-        level = run._levels[0]
-        uploads = {id(a[0]): a[1] for name, a in log if name == "upload"}
-        self.assertTrue(np.isnan(uploads[id(level.spare)]).all())
-        self.assertTrue(np.isfinite(level.layout.to_dense(uploads[id(level.current)])).all())
-
-    def test_launch_order_per_step(self):
-        for objects, order in (((), ["BlockHaloCopy", "BlockStep"]),
-                               ((circle(),), ["BlockHaloCopy", "BlockCurvedBoundary", "BlockStep"])):
-            with self.subTest(bodies=len(objects)):
-                run, log, _ = backend(objects)
-                log.clear()
-                run.step(2)
-                self.assertEqual([name for name, _ in log], order * 2 + ["sync"])
-                step = [a for name, a in log if name == "BlockStep"][0]
-                self.assertEqual(int(step[4]), 0)  # null kind: every core cell ACTIVE
-
-    def test_circle_solids_are_not_state(self):
-        run, _, _ = backend((circle(),))
-        solid = run._device_boundaries[0].tables["solid"]
-        self.assertTrue(solid.any())
-        np.testing.assert_array_equal(run._state[0], ~solid)
-
-    def test_initialize_upload_failure_fails_instance(self):
-        run, _, host = backend(initialize=False)
-        host.cuda.memcpy_htod.side_effect = RuntimeError("upload failed")
-        with self.assertRaisesRegex(RuntimeError, "upload failed"):
-            run.initialize(state())
-        with self.assertRaisesRegex(RuntimeError, "failed"):
-            run.step(1)
-
-    def test_launch_failure_fails_instance_and_keeps_step_count(self):
-        for kernel, objects in (("BlockHaloCopy", ()), ("BlockCurvedBoundary", (circle(),)),
-                                ("BlockStep", ())):
-            with self.subTest(kernel=kernel):
-                run, log, host = backend(objects, failing_kernel=kernel)
-                run.step(1)
-                host.armed = True
-                log.clear()
-                with self.assertRaisesRegex(RuntimeError, kernel + " launch failed"):
-                    run.step(2)
-                self.assertEqual(run.step_count, 1)
-                self.assertNotIn("sync", [name for name, _ in log])
-                with self.assertRaisesRegex(RuntimeError, "failed"):
-                    run.download()
-
-    def test_free_memory_preflight_before_stream_and_buffers(self):
-        for cfg, options in ((LBMConfig(nx=32, ny=24, viscosity=.1),
-                              dict(refinements=[], level_blocks=[(12, 16)])),
-                             (LBMConfig(nx=48, ny=48, viscosity=.02),  # every level counts
-                              dict(refinements=[[(6, 6, 26, 26)]], level_blocks=[(4, 4)] * 2))):
-            with self.subTest(depth=len(options["refinements"]) + 1):
-                host = host_module()
-                host.cuda.mem_get_info.return_value = (1024, 1 << 30)
-                with self.assertRaisesRegex(MemoryError, "exceed free device memory"):
-                    host.FixedPatchStepper(cfg, context=Mock(activate=nullcontext, sm_arch="sm_70"),
-                                           **options)
-                host.cuda.Stream.assert_not_called()
-                host.cuda.mem_alloc.assert_not_called()
-                host.compiler.build_module.assert_not_called()
-
-    def test_wait_failure_after_a_good_step_keeps_its_count(self):
-        run, _, _ = backend()
-        run.step(1)
-        run._stream.synchronize.side_effect = RuntimeError("device fault")
-        with self.assertRaises(RuntimeError):
-            run.step(2)
-        self.assertEqual((run.step_count, run._status), (1, "failed"))
-
-    def test_final_wait_failure_fails_instance_and_keeps_step_count(self):
-        run, _, _ = backend()
-        run._stream.synchronize.side_effect = RuntimeError("device error")
-        with self.assertRaisesRegex(RuntimeError, "device error"):
-            run.step(3)
-        self.assertEqual(run.step_count, 0)
-        with self.assertRaisesRegex(RuntimeError, "failed"):
-            run.step(1)
-
-    def test_step_n_waits_once_and_matches_n_single_steps(self):
-        many, log_many, _ = backend()
-        many.step(3)
-        single, log_single, _ = backend()
-        for _ in range(3):
-            single.step(1)
-        names = [name for name, _ in log_many]
-        self.assertEqual(names.count("sync"), 1)
-        self.assertEqual(names[-1], "sync")
-        self.assertEqual(names.count("BlockStep"), 3)
-
-        def roles(run, log):  # same launch order and buffer roles as three step(1) calls
-            level = run._levels[0]
-            tags = {id(level.current): "c", id(level.spare): "s"}
-            return [(n, tuple(tags.get(id(x), "-") for x in a)) for n, a in log if n != "sync"]
-        self.assertEqual(roles(many, log_many), roles(single, log_single))
-        self.assertIsNot(many._levels[0].current, many._levels[0].spare)
-        self.assertEqual((many.step_count, single.step_count), (3, 3))
-
-    def test_step_zero_launches_and_waits_for_nothing(self):
-        run, log, _ = backend()
-        log.clear()
-        run.step(0)
-        self.assertEqual((log, run.step_count), ([], 0))
-
-    def test_launch_failure_mid_batch_fails_without_waiting(self):
-        run, log, host = backend()
-        step = run._levels[0].kernel
-        calls = []
-
-        def failing(*args, **kwargs):
-            calls.append(1)
-            if len(calls) == 2:
-                raise RuntimeError("launch failure")
-            return step(*args, **kwargs)
-
-        run._levels[0].kernel = failing
-        log.clear()
-        with self.assertRaises(RuntimeError):
-            run.step(3)
-        self.assertNotIn("sync", [name for name, _ in log])
-        self.assertEqual((run.step_count, run._status), (0, "failed"))
-        with self.assertRaises(RuntimeError):
-            run.download()
-
-    def test_initialize_rejects_bad_state_before_upload(self):
-        run, log, _ = backend(initialize=False)
-        good = state()
-        for bad in (good[:, :-1], np.where(np.arange(32) == 3, np.nan, good), -good):
-            with self.assertRaises(ValueError):
-                run.initialize(bad)
-        self.assertNotIn("upload", [name for name, _ in log])
-        run.initialize(good)
-        self.assertEqual([name for name, _ in log].count("upload"), 2)
-
-    def test_download_failure_fails_instance(self):
-        run, _, host = backend()
-        host.cuda.memcpy_dtoh.side_effect = RuntimeError("read failed")
-        with self.assertRaisesRegex(RuntimeError, "read failed"):
-            run.download()
-        for operation in (run.download, run.step):
-            with self.assertRaisesRegex(RuntimeError, "failed"):
-                operation()
-
-    def test_activation_failure_retains_resources_for_retry(self):
-        run, _, _ = backend()
-        allocations, stream = run._allocations[:], run._stream
-        unavailable = MagicMock()
-        unavailable.__enter__.side_effect = RuntimeError("activation failed")
-        run.ctx.activate = Mock(side_effect=[unavailable, nullcontext()])
-        with self.assertRaisesRegex(RuntimeError, "activation failed"):
-            run.close()
-        self.assertEqual(run._status, "failed")
-        self.assertEqual(run._allocations, allocations)
-        self.assertIs(run._stream, stream)
-        run.close()
-        self.assertEqual(run._status, "closed")
-        self.assertEqual(run._levels, [])
-        for allocation in allocations:
-            allocation.free.assert_called_once()
-
-
-if __name__ == "__main__":
-    unittest.main()
@@ -295,7 +295,6 @@ def test_passive_overlap_failure_keeps_live_state(device, entry, second_x):
                          _initialized=entry == "runtime_edit",
                          _assert_allocation_contract=lambda: None,
                          stream=SimpleNamespace(synchronize=device.synchronize))
-    sim._require_usable = Simulation._require_usable.__get__(sim)
    with pytest.raises(ValueError, match="active fluid receivers|unique cut-link writers"):
        if entry == "sync":
            bodies.sync_to_gpu(field)
@@ -1,170 +0,0 @@
-"""Host contract of FixedPatchStepper.step(n): one stream wait per call (mocked driver).
-
-All n base intervals are enqueued in order on the backend stream and the host
-waits once before step returns. Launch order and buffer arguments must equal n
-step(1) calls; completion metadata, raw-exchange and window ledgers, and failure
-invalidation are published only after that wait.
-"""
-from contextlib import nullcontext
-from unittest.mock import Mock
-import unittest
-
-import numpy as np
-
-from CelerisLab.config import LBMConfig
-from tests.unit.test_raw_exchange import body
-from tests.unit.test_refinement_three_level_multipatch import host_module
-
-BODY_LEVEL, HOPS = 3, 8  # body on L3: 2**3 hops per base interval
-
-
-def backend(window=False):
-    """Four-level host backend; every launch, memset and stream wait goes to one log."""
-    host = host_module()
-    log, allocations = [], []
-
-    def allocate(size):
-        allocations.append(Mock(size=size))
-        return allocations[-1]
-
-    module = Mock()
-    module.get_global.return_value = (Mock(), 20)
-    module.get_function.side_effect = lambda name: Mock(
-        side_effect=lambda *a, **k: log.append((name, a, k)))
-    host.compiler.build_module.return_value = module
-    host.cuda.mem_alloc.side_effect = allocate
-    host.cuda.memcpy_dtoh.side_effect = lambda array, pointer: array.fill(0)
-    host.cuda.memset_d8_async.side_effect = lambda *a: log.append(("memset", a, {}))
-    run = host.FixedPatchStepper(
-        LBMConfig(nx=48, ny=48, viscosity=.02), context=Mock(activate=nullcontext, sm_arch="sm_70"),
-        refinements=[[(6, 6, 26, 26)], [(20, 20, 44, 44)], [(56, 56, 72, 72)]],
-        level_blocks=[(4, 4)] * 4, level_objects=[(), (), (), (body(),)],
-        observe_raw_exchange=True, observe_window_exchange=window)
-    run._stream.synchronize.side_effect = lambda: log.append(("sync", (), {}))
-    run.initialize(*[np.full(level.shape, 1 / 9, np.float32) for level in run._levels])
-    log.clear()
-    return run, log, allocations
-
-
-def signature(log, allocations):
-    """Event names with device-pointer arguments replaced by allocation index."""
-    index = {id(a): i for i, a in enumerate(allocations)}
-    return [(name, tuple(index.get(id(a), "-") for a in args)) for name, args, _ in log]
-
-
-def names(log):
-    return [name for name, _, _ in log]
-
-
-class StepSync(unittest.TestCase):
-    def test_step1_waits_once_after_all_launches(self):
-        run, log, _ = backend()
-        run.step(1)
-        self.assertEqual(names(log).count("sync"), 1)
-        self.assertEqual(names(log)[-1], "sync")
-        self.assertEqual(names(log)[0], "memset")
-        self.assertEqual(names(log).count("BlockStep"), 1 + 2 + 4 + 8)
-        self.assertEqual(run.step_count, 1)
-        run.close()
-
-    def test_step_n_equals_n_step1_launches_with_one_wait(self):
-        single, single_log, single_alloc = backend()
-        for _ in range(3):
-            single.step(1)
-        batched, batched_log, batched_alloc = backend()
-        batched.step(3)
-        expected = [e for e in signature(single_log, single_alloc) if e[0] != "sync"]
-        got = signature(batched_log, batched_alloc)
-        self.assertEqual(got[-1][0], "sync")
-        self.assertEqual(names(batched_log).count("sync"), 1)
-        self.assertEqual(got[:-1], expected)  # same order, same ping-pong buffers
-        self.assertEqual(names(batched_log).count("memset"), 3)
-        self.assertEqual(batched.step_count, single.step_count)
-        index = [{id(a): i for i, a in enumerate(alloc)} for alloc in (single_alloc, batched_alloc)]
-        self.assertEqual(index[1][id(batched._levels[0].current)],
-                         index[0][id(single._levels[0].current)])
-        single.close()
-        batched.close()
-
-    def test_step0_launches_and_waits_for_nothing(self):
-        run, log, _ = backend()
-        run.step(2)
-        log.clear()
-        run.step(0)
-        self.assertEqual(log, [])
-        self.assertEqual(run.step_count, 2)
-        value = run.read_raw_exchange(0, level=BODY_LEVEL)  # step(0) keeps the ledger valid
-        self.assertEqual((value["time_start"], value["time_end"]), (1, 2))
-        run.close()
-
-    def test_completion_metadata_covers_all_intervals(self):
-        run, _, _ = backend()
-        with self.assertRaisesRegex(RuntimeError, "completed"):
-            run.read_raw_exchange(0, level=BODY_LEVEL)
-        run.step(3)
-        value = run.read_raw_exchange(0, level=BODY_LEVEL)
-        self.assertEqual((value["time_start"], value["time_end"], value["sample_count"]),
-                         (2, 3, HOPS))  # step(n) retains only the final base interval
-        run.step(2)
-        self.assertEqual(run.read_raw_exchange(0, level=BODY_LEVEL)["time_end"], 5)
-        run.close()
-
-    def test_window_accounting_across_batched_calls(self):
-        windows = []
-        for calls in ((1, 3), (4,), (1, 1, 1, 1)):
-            run, log, _ = backend(window=True)
-            run.step(1)
-            run.start_exchange_window()
-            log.clear()
-            for n in calls:
-                run.step(n)
-            self.assertEqual(names(log).count("BlockAccumulateExchange"), 2 * 4 * HOPS)
-            self.assertEqual(names(log).count("sync"), len(calls))
-            out = run.read_exchange_window(0, level=BODY_LEVEL)
-            windows.append((out["time_start"], out["time_end"], out["sample_count"]))
-            run.close()
-        self.assertEqual(windows, [(1., 5, 4 * HOPS)] * 3)
-
-    def test_failed_wait_publishes_nothing(self):
-        run, _, _ = backend(window=True)
-        run.step(2)
-        run.start_exchange_window()
-        run._stream.synchronize.side_effect = RuntimeError("device failure")
-        with self.assertRaisesRegex(RuntimeError, "device failure"):
-            run.step(3)
-        self.assertEqual(run.step_count, 2)  # value at entry to the failing call
-        self.assertFalse(run._exchange_valid)
-        self.assertFalse(run._window_valid)
-        for read in (lambda: run.read_raw_exchange(0, level=BODY_LEVEL),
-                     lambda: run.read_exchange_window(0, level=BODY_LEVEL), run.download,
-                     lambda: run.step(1)):
-            with self.assertRaisesRegex(RuntimeError, "failed"):
-                read()
-        run._stream.synchronize.side_effect = None
-        run.close()
-
-    def test_launch_error_mid_batch_skips_wait_and_fails(self):
-        run, log, _ = backend()
-        run.step(1)
-        log.clear()
-        hop, calls = run._hop, []
-
-        def failing(index, *args, **kw):
-            if index == 0:
-                calls.append(1)
-                if len(calls) == 2:
-                    raise RuntimeError("launch failure")
-            return hop(index, *args, **kw)
-        run._hop = failing
-        with self.assertRaisesRegex(RuntimeError, "launch failure"):
-            run.step(3)
-        self.assertNotIn("sync", names(log))
-        self.assertEqual(run.step_count, 1)
-        self.assertFalse(run._exchange_valid)
-        with self.assertRaisesRegex(RuntimeError, "failed"):
-            run.read_raw_exchange(0, level=BODY_LEVEL)
-        run.close()
-
-
-if __name__ == "__main__":
-    unittest.main()
@@ -12,15 +12,13 @@ from source_provenance import checkout_provenance
 import re100_helpers
 import run_jet_circle_cv

-sys.path.insert(0, str(ROOT / "tests/unit"))
-from historical_tree import historical_tree, tracked
-
 ORIGIN = "47b05bc8abfc957fe876b0c63bdb386e3c6cbaa5"


-def test_manifests_cover_current_source_and_preserve_origin(tmp_path):
-    assert re100_helpers.PIN == run_jet_circle_cv.PIN == ORIGIN
-    manifests = {}
+def test_manifests_cover_current_source_and_preserve_origin():
+    names = subprocess.check_output(
+        ["git", "-C", str(ROOT), "ls-files", "src"], text=True
+    ).splitlines()
    for group, hashes_key, revision_key in (
        ("dfg_re100", "production_sha256", "commit"),
        ("jet_circle_cv", "files", "revision"),
@@ -31,22 +29,18 @@ def test_manifests_cover_current_source_and_preserve_origin(tmp_path):
            ).read_text()
        )
        assert manifest[revision_key] == ORIGIN
+        assert set(manifest[hashes_key]) == set(names)
        assert manifest["source_origin_commits"][-1] == ORIGIN
        assert (
            manifest["historical_checkpoints"]["production"]
            == "d79e3f6f0a424f5bcc3a1a6d610313e49daf97a2"
        )
        assert "not inherited from d79" in manifest["experiment"]
-        manifests[group] = manifest[hashes_key]
-    # History-dependent checks last, so a missing commit masks nothing above.
-    names = tracked(ORIGIN)
-    origin = historical_tree(ORIGIN, tmp_path)
-    for group, hashes in manifests.items():
-        assert set(hashes) == set(names), group
-        for name, expected in hashes.items():
+        for name, expected in manifest[hashes_key].items():
            assert (
-                hashlib.sha256((origin / name).read_bytes()).hexdigest() == expected
+                hashlib.sha256((ROOT / name).read_bytes()).hexdigest() == expected
            ), name
+    assert re100_helpers.PIN == run_jet_circle_cv.PIN == ORIGIN


 def test_runtime_checkout_identity_is_separate_from_source_origin(tmp_path):
@@ -0,0 +1,75 @@
+"""Per-hop mass ledger of a fluid cell set A (host-only, stdlib, D2Q9 pull scheme).
+
+Diagnostic definition only; no solver code, no GPU. Arrays are indexed [q][y][x] on a
+periodic nx x ny lattice; A is a set of (x, y) ACTIVE cells.
+
+  O  owner populations before any staging (canonical owner slots, start of the hop)
+  P  populations each cell actually pulls: P[q](x) is what x reads for direction q
+     (after halo copy, BFL/coalesce staging, ...)
+  D  populations written by the cell after collision
+
+With M0 = sum_{A,q} O and M1 = sum_{A,q} D (both owner snapshots of A):
+
+  M1 - M0 + Q_exact = C + H                                            (exact)
+
+  C       = sum_{x in A, q} (D_q(x) - P_q(x))              collision source
+  H       = sum_{x in A, q} (P_q(x) - O_q(x - c_q))        actual minus ideal owner input,
+            split into H_seam (x - c_q in A: internal pulls, e.g. tile seams) and
+            H_edge (x - c_q not in A: inputs crossing the edge of A)
+  Q_exact = sum_{y in A, y+c_q not in A} O_q(y) - sum_{y not in A, y+c_q in A} O_q(y)
+            owner-based net outward flux (outgoing minus incoming)
+
+Proof: sum_{A,q} O_q(x - c_q) - M0 = -Q_exact (relabel y = x - c_q per direction), and
+sum_{A,q} D = C + H + sum_{A,q} O_q(x - c_q). Q_exact must use O, not staged values: a
+flux built from staged inputs already contains H_edge, so adding the full H counts it
+twice (see staged_flux). The edge of A must not cross cut links, otherwise O at the solid
+end enters Q_exact.
+"""
+
+CX = (0, 1, -1, 0, 0, 1, -1, 1, -1)
+CY = (0, 0, 0, 1, -1, 1, -1, -1, 1)
+
+
+def _src(x, y, q, nx, ny):
+    return (x - CX[q]) % nx, (y - CY[q]) % ny
+
+
+def pull(O, nx, ny):
+    """Ideal pulled input: P[q](x) = O[q](x - c_q)."""
+    return [[[O[q][_src(x, y, q, nx, ny)[1]][_src(x, y, q, nx, ny)[0]]
+              for x in range(nx)] for y in range(ny)] for q in range(9)]
+
+
+def ledger(O, P, D, A, nx, ny):
+    A = set(A)
+    out = dict(dM=0.0, C=0.0, H_seam=0.0, H_edge=0.0, Q_exact=0.0)
+    for (x, y) in A:
+        for q in range(9):
+            out["dM"] += D[q][y][x] - O[q][y][x]
+            out["C"] += D[q][y][x] - P[q][y][x]
+            sx, sy = _src(x, y, q, nx, ny)
+            key = "H_seam" if (sx, sy) in A else "H_edge"
+            out[key] += P[q][y][x] - O[q][sy][sx]
+            if (sx, sy) not in A:
+                out["Q_exact"] -= O[q][sy][sx]          # incoming owner population
+            tx, ty = (x + CX[q]) % nx, (y + CY[q]) % ny
+            if (tx, ty) not in A:
+                out["Q_exact"] += O[q][y][x]            # outgoing owner population
+    out["H"] = out["H_seam"] + out["H_edge"]
+    out["residual"] = out["dM"] + out["Q_exact"] - out["C"] - out["H"]
+    return out
+
+
+def staged_flux(O, P, A, nx, ny):
+    """Outward flux with incoming terms from the staged inputs P (the wrong combination)."""
+    A = set(A)
+    q_out = 0.0
+    for (x, y) in A:
+        for q in range(9):
+            sx, sy = _src(x, y, q, nx, ny)
+            if (sx, sy) not in A:
+                q_out -= P[q][y][x]
+            tx, ty = (x + CX[q]) % nx, (y + CY[q]) % ny
+            if (tx, ty) not in A:
+                q_out += O[q][y][x]
+    return q_out
@@ -30,7 +30,7 @@ original_module = gate.compile_module
 cache = {}
 def retain_compile(*args, **kwargs):
    result = original_compile(*args, **kwargs)
-    source = kwargs.get('source', 'uniform.cu')
+    source = kwargs.get('source', 'kernel_v2.cu')
    shutil.copyfile(result, a.output / (Path(source).stem + '.ptx'))
    metadata = Path(result + '.json')
    if metadata.exists(): shutil.copyfile(metadata, a.output / (Path(source).stem + '.ptx.json'))
@@ -1,4 +1,4 @@
-"""Single blocked level, FixedPatchStepper(refinements=[]): CPU contracts plus short GPU cases.
+"""Periodic body-free block backend: CPU contract doubles plus one short GPU case.

 Frozen GPU gates (32x24, 8 steps, positive low-Mach non-equilibrium state):
 finite cores; bit-identical 1x1, 2x2 and 4x2 splits; max |G - FP64 oracle|
@@ -11,8 +11,9 @@ Mass/momentum drift is reported, not gated.
 from contextlib import nullcontext
 from pathlib import Path
 import resource
+from threading import RLock
 import unittest
-from unittest.mock import Mock, patch
+from unittest.mock import MagicMock, Mock, patch

 import numpy as np
 import pycuda.driver as cuda
@@ -20,10 +21,10 @@ import pycuda.driver as cuda
 from CelerisLab.body.geometry.circle import CircleGeometry
 from CelerisLab.body.objects import SimObject
 from CelerisLab.config import LBMConfig
-from CelerisLab.cuda import compiler
-from CelerisLab.lbm import refinement_stepper as implementation
-from CelerisLab.lbm.blocks import NO_DONOR, BlockLayout2D, static_circle_tables
-from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
+from CelerisLab.cuda import compiler_v2 as compiler
+from CelerisLab.lbm import block_stepper as implementation
+from CelerisLab.lbm.blocks import NO_DONOR, BlockLayout2D
+from CelerisLab.lbm.block_stepper import PeriodicBlockStepper
 from CelerisLab.lbm.descriptors import D2Q9_EX, D2Q9_EY

 ROOT = Path(__file__).resolve().parents[3]
@@ -111,78 +112,72 @@ def totals(g):
    return g.sum(), (CX[:, None, None] * g).sum(), (CY[:, None, None] * g).sum()


-def single(cfg=None, *, context, block=(12, 16), objects=(), **kw):
-    """One periodic blocked level (the former PeriodicBlockStepper capability)."""
-    return FixedPatchStepper(cfg or config(), context=context, refinements=[], level_blocks=[block],
-                             level_objects=[tuple(objects)], **kw)
-
-
 class CpuContracts(unittest.TestCase):
-    """No device: driver doubles prove rejection before allocation."""
+    """No device: driver doubles prove rejection before allocation and lifecycle."""

    def forbidden_context(self):
        return Mock(activate=Mock(side_effect=AssertionError("CUDA reached")), sm_arch="sm_70")

    def test_unsupported_inputs_rejected_before_cuda(self):
-        for cfg, kw in ((None, dict(block=None)),
-                        (None, dict(coarse_boundary="channel")),
-                        (None, dict(coarse_boundary="slot_reservoir")),
-                        (config(collision="MRT"), {}),
-                        (config(les_enabled=True), {}),
-                        (None, dict(explosion="projected_linear")),
-                        (None, dict(observe_raw_exchange=True)),
-                        (None, dict(observe_raw_exchange=True, observe_window_exchange=True)),
-                        (None, dict(level_jets=[(object(),)]))):
-            with self.subTest(kw=kw), self.assertRaises(ValueError):
-                single(cfg, context=self.forbidden_context(), **kw)
+        layout = BlockLayout2D.build(NY, NX, 12, 16)
+        cases = ((config(), BlockLayout2D.build(NY, NX, 12, 16, periodic=(True, False))),
+                 (config(), BlockLayout2D.build(NY, NX * 2, 12, 16)),
+                 (config(collision="MRT"), layout),
+                 (config(), "layout"))
+        for cfg, lay in cases:
+            with self.subTest(cfg=cfg.collision, layout=getattr(lay, "periodic", lay)):
+                with self.assertRaises(ValueError):
+                    PeriodicBlockStepper(cfg, lay, context=self.forbidden_context())
        moving = circle()
        moving.state.vx = .01
        build_links = CircleGeometry.build_cut_links
        for objects in ([moving], [circle((1.5, 11.5), 2.0)], [circle(), circle()]):
            with self.assertRaises(ValueError):
-                single(context=self.forbidden_context(), objects=objects)
+                PeriodicBlockStepper(config(), layout, context=self.forbidden_context(), objects=objects)
        for center, radius in (((16.0, 12.0), -.1), ((16.0, 12.0), -.5), ((15.5, 11.5), np.nan),
                               ((np.inf, 11.5), 4.3), ((15.5, np.nan), 4.3)):
            with self.subTest(center=center, radius=radius), self.assertRaises(ValueError):
-                single(context=self.forbidden_context(), objects=[circle(center, radius)])
-        with patch.object(implementation, "static_circle_tables") as prepare:
+                PeriodicBlockStepper(config(), layout, context=self.forbidden_context(),
+                                     objects=[circle(center, radius)])
+        with patch.object(implementation, "_static_circles") as prepare:
            with self.assertRaisesRegex(AssertionError, "CUDA reached"):  # body-free path
-                single(context=self.forbidden_context(), objects=iter(()))
+                PeriodicBlockStepper(config(), layout, context=self.forbidden_context(), objects=iter(()))
            prepare.assert_not_called()
-        with self.assertRaisesRegex(AssertionError, "CUDA reached"):
-            single(context=self.forbidden_context(), objects=[circle()])
        with patch.object(CircleGeometry, "build_cut_links",
                          lambda geometry, *a, **k: build_links(geometry, *a, **k)[1:]):
            with self.assertRaisesRegex(ValueError, "cover every fluid pull"):
-                single(context=self.forbidden_context(), objects=[circle()])
+                PeriodicBlockStepper(config(), layout, context=self.forbidden_context(),
+                                     objects=[circle()])
        with patch.object(implementation, "MAX_GRID_DIM", 3):
            with self.assertRaisesRegex(ValueError, "launch limits"):
-                single(context=self.forbidden_context())
+                PeriodicBlockStepper(config(), layout, context=self.forbidden_context())

    def test_memory_and_param_layout_rejected_before_allocation(self):
+        layout = BlockLayout2D.build(NY, NX, 12, 16)
        context = Mock(activate=nullcontext, sm_arch="sm_70")
        with patch.object(cuda, "mem_get_info", return_value=(1024, 1 << 30)), \
                patch.object(cuda, "Stream") as stream, patch.object(cuda, "mem_alloc") as alloc:
-            with self.assertRaisesRegex(MemoryError, "exceed free device memory"):
-                single(context=context, objects=[circle()])
+            with self.assertRaises(MemoryError):
+                PeriodicBlockStepper(config(), layout, context=context)
            stream.assert_not_called()
            alloc.assert_not_called()
        module = Mock()
        module.get_global.return_value = (Mock(), 16)
-        with patch.object(cuda, "Stream"), patch.object(cuda, "mem_get_info", return_value=(1 << 40, 1 << 40)), patch.object(compiler, "generate_config"), \
+        with patch.object(cuda, "mem_get_info", return_value=(1 << 30, 1 << 30)), \
+                patch.object(cuda, "Stream"), patch.object(compiler, "generate_config"), \
                patch.object(compiler, "compile_kernel") as compile_kernel, \
                patch.object(compiler, "cleanup_build_dir"), \
                patch.object(compiler, "load_module", return_value=module), \
                patch.object(cuda, "mem_alloc") as alloc, patch.object(cuda, "memcpy_htod") as upload:
            with self.assertRaisesRegex(RuntimeError, "layout mismatch"):
-                single(context=context)
-            self.assertEqual(compile_kernel.call_args.kwargs["source"], "blocked.cu")
+                PeriodicBlockStepper(config(), layout, context=context)
+            self.assertEqual(compile_kernel.call_args.kwargs["source"], "blocks.cu")
            alloc.assert_not_called()
            upload.assert_not_called()

    def test_static_circle_tables(self):
        layout = BlockLayout2D.build(NY, NX, 12, 16)
-        tables = static_circle_tables(layout, (circle(),))
+        tables = implementation._static_circles(layout, (circle(),))
        links = CircleGeometry(*CENTER, RADIUS).build_cut_links(NX, NY, domain_flags=dense_flags())
        self.assertEqual(len(tables["cell"]), len(links))
        np.testing.assert_array_equal(tables["q"], np.array([l.q for l in links], np.float32))
@@ -190,11 +185,84 @@ class CpuContracts(unittest.TestCase):
        self.assertTrue((tables["flags"][solid_cells] & 2).all())
        self.assertEqual(int(np.count_nonzero(tables["flags"] & 2)), int(tables["solid"].sum()))

+    def fake_backend(self):
+        backend = PeriodicBlockStepper.__new__(PeriodicBlockStepper)
+        backend.layout = BlockLayout2D.build(NY, NX, 12, 16)
+        backend.ctx = Mock(activate=nullcontext)
+        backend._lock = RLock()
+        backend._status = "ready"
+        backend._step_count = 0
+        backend._stream = Mock()
+        backend._module = Mock()
+        backend._copy = backend._step = Mock()
+        backend.boundary, backend._curved, backend._tables, backend._links = None, None, {}, 0
+        backend._dst, backend._src = Mock(), Mock()
+        backend._block, backend._grid, backend._copy_grid = (32, 1, 1), (1, 1, 1), (1, 1)
+        backend._allocations = [Mock(), Mock()]
+        backend._current, backend._spare = backend._allocations
+        return backend

-class SingleLevelGPU(unittest.TestCase):
+    def test_initialize_rejects_bad_state_before_upload(self):
+        backend = self.fake_backend()
+        good = initial_state()
+        bad = (good[:, :-1], np.where(np.arange(NX) == 3, np.nan, good), -good)
+        with patch.object(cuda, "memcpy_htod") as upload:
+            for g in bad:
+                with self.assertRaises(ValueError):
+                    backend.initialize(g)
+            upload.assert_not_called()
+            backend.initialize(good)
+            self.assertEqual(upload.call_count, 2)
+            self.assertTrue(np.isnan(upload.call_args_list[1].args[1]).all())
+        backend.close()
+
+    def test_failed_launch_and_download_make_backend_unusable(self):
+        for failure in ("step", "download"):
+            backend = self.fake_backend()
+            if failure == "step":
+                backend._step = Mock(side_effect=RuntimeError("launch failed"))
+                with self.assertRaisesRegex(RuntimeError, "launch failed"):
+                    backend.step()
+            else:
+                with patch.object(cuda, "memcpy_dtoh", side_effect=RuntimeError("read failed")):
+                    with self.assertRaisesRegex(RuntimeError, "read failed"):
+                        backend.download()
+            for operation in (backend.download, backend.step):
+                with self.assertRaisesRegex(RuntimeError, "failed"):
+                    operation()
+            backend.close()
+            for allocation in backend_allocations(backend):
+                allocation.free.assert_called_once()
+
+    def test_activation_failure_retains_resources_for_retry(self):
+        backend = self.fake_backend()
+        allocations, stream, module = backend._allocations[:], backend._stream, backend._module
+        unavailable = MagicMock()
+        unavailable.__enter__.side_effect = RuntimeError("activation failed")
+        backend.ctx.activate = Mock(side_effect=[unavailable, nullcontext()])
+        with self.assertRaisesRegex(RuntimeError, "activation failed"):
+            backend.close()
+        self.assertEqual(backend._status, "failed")
+        self.assertEqual(backend._allocations, allocations)
+        self.assertIs(backend._stream, stream)
+        self.assertIs(backend._module, module)
+        backend.close()
+        self.assertEqual(backend._status, "closed")
+        self.assertIsNone(backend._module)
+        self.assertIsNone(backend._copy)
+        self.assertIsNone(backend._step)
+        for allocation in allocations:
+            allocation.free.assert_called_once()
+
+
+def backend_allocations(backend):
+    return (backend._current, backend._spare)
+
+
+class PeriodicBlockGPU(unittest.TestCase):
    def setUp(self):
        self.assertEqual(Path(implementation.__file__).resolve(),
-                         ROOT / "src/CelerisLab/lbm/refinement_stepper.py")
+                         ROOT / "src/CelerisLab/lbm/block_stepper.py")

    def tearDown(self):
        rss = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss / 1024
@@ -205,13 +273,13 @@ class SingleLevelGPU(unittest.TestCase):
        g0 = initial_state()
        results = []
        with CudaContext(0) as context:
-            for split in SPLITS:
-                with single(context=context, block=split) as backend:
+            for by, bx in SPLITS:
+                layout = BlockLayout2D.build(NY, NX, by, bx)
+                with PeriodicBlockStepper(config(), layout, context=context) as backend:
                    backend.initialize(g0)
                    backend.step(STEPS)
                    self.assertEqual(backend.step_count, STEPS)
-                    (g,) = backend.download()
-                    results.append(g)
+                    results.append(backend.download())
        for (by, bx), g in zip(SPLITS, results):
            self.assertTrue(np.isfinite(g).all(), (by, bx))
            np.testing.assert_array_equal(g, results[0], err_msg=f"split {(by, bx)}")
@@ -238,17 +306,16 @@ class SingleLevelGPU(unittest.TestCase):
        g0 = initial_state()
        results = []
        with CudaContext(0) as context:
-            for split in SPLITS:
-                with single(context=context, block=split, objects=[circle()]) as backend:
-                    tables = backend._device_boundaries[0].tables
+            for by, bx in SPLITS:
+                layout = BlockLayout2D.build(NY, NX, by, bx)
+                with PeriodicBlockStepper(config(), layout, context=context, objects=[circle()]) as backend:
                    for key, values, dtype in (("q", [l.q for l in links], np.float32),
                                               ("dir", [l.dir for l in links], np.uint8),
                                               ("fallback", [l.fallback for l in links], np.uint8)):
-                        np.testing.assert_array_equal(tables[key], np.array(values, dtype))
+                        np.testing.assert_array_equal(backend.boundary[key], np.array(values, dtype))
                    backend.initialize(g0)
                    backend.step(STEPS)
-                    (g,) = backend.download()
-                    results.append(g)
+                    results.append(backend.download())
        for (by, bx), g in zip(SPLITS, results):
            self.assertTrue(np.isfinite(g[:, ~solid]).all(), (by, bx))
            self.assertTrue(np.isnan(g[:, solid]).all(), (by, bx))
@@ -11,7 +11,7 @@ import numpy as np
 import pycuda.driver as cuda

 from CelerisLab.config import LBMConfig
-from CelerisLab.cuda import compiler
+from CelerisLab.cuda import compiler_v2 as compiler
 from CelerisLab.cuda.context import CudaContext
 from CelerisLab.lbm.blocks import BlockLayout2D
 from CelerisLab.lbm.descriptors import FLUID, SOLID, BC_WALL, BC_INLET, BC_OUTLET
@@ -27,7 +27,7 @@ def state(ny, nx):

 def compile_module(cfg, context, source):
    cfg.validate()
-    build = compiler.generate_config(cfg, block_root_bc=2 if source == "blocked.cu" else 0)
+    build = compiler.generate_config(cfg)
    try:
        module = compiler.load_module(compiler.compile_kernel(context.sm_arch, source=source, build_dir=build))
    finally:
@@ -46,7 +46,7 @@ def run_root(cfg, context, initial, block=None, steps=8):
    try:
        layout = None if block is None else BlockLayout2D.build(cfg.ny, cfg.nx, *block)
        local = cfg if layout is None else replace(cfg, nx=layout.local_shape[1], ny=layout.local_shape[0])
-        module = compile_module(local, context, "uniform.cu" if layout is None else "blocked.cu")
+        module = compile_module(local, context, "kernel_v2.cu" if layout is None else "blocks.cu")
        data = initial if layout is None else layout.to_blocked(initial)
        current, spare = upload(data), upload(np.full_like(data, np.nan))
        tpb = cfg.threads_per_block
@@ -106,10 +106,9 @@ class ChannelRootGPU(unittest.TestCase):
                    cfg = LBMConfig(nx=48, ny=32, viscosity=.1, velocity=.015, ddf_shifting=shifted)
                    reference = None
                    for block in ((32,48), (8,8), (4,1)):
-                        with FixedPatchStepper(cfg, context=context, coarse_boundary="channel",
-                                refinements=[(8,6,40,26)], level_blocks=[block, (8,8)],
-                                level_objects=[(), (body,) if bodies else ()],
-                                level_jets=[(), (jet,) if bodies else ()]) as backend:
+                        with FixedPatchStepper(cfg, (8,6,40,26), context=context,
+                                coarse_boundary="channel", base_block=block, fine_block=(8,8),
+                                fine_objects=(body,) if bodies else (), jets=(jet,) if bodies else ()) as backend:
                            arrays = [state(*level.shape[1:]) for level in backend._levels]
                            backend.initialize(*arrays)
                            backend.step(4)
@@ -25,8 +25,8 @@ EPS = np.finfo(np.float64).eps


 def build(m, window=True, raw=True):
-    return FixedPatchStepper(config("periodic"), context=m["context"], refinements=[BOUNDS],
-                             level_blocks=[(3, 4), (16, 20)], level_objects=[(), (circle(),)],
+    return FixedPatchStepper(config("periodic"), BOUNDS, context=m["context"], base_block=(3, 4),
+                             fine_block=(16, 20), fine_objects=[circle()],
                             observe_raw_exchange=raw, observe_window_exchange=window)


@@ -59,7 +59,7 @@ class ExchangeWindowCPU(unittest.TestCase):
            stack.enter_context(patch.object(cuda, "memcpy_dtoh",
                                             side_effect=lambda out, src: out.fill(0)))
            backend = build(m)
-            b = backend._device_boundaries[1]
+            b = backend._level_boundaries[1]
            self.assertIsNotNone(b.window_integral)
            with self.assertRaises(RuntimeError):
                backend.start_exchange_window()  # not initialized
@@ -96,7 +96,7 @@ class ExchangeWindowCPU(unittest.TestCase):
        with mocked_backend() as m, ExitStack() as stack:
            stack.enter_context(patch.object(cuda, "memset_d8_async"))
            backend = build(m, window=False)
-            self.assertIsNone(backend._device_boundaries[1].window_integral)
+            self.assertIsNone(backend._level_boundaries[1].window_integral)
            backend.initialize(*[np.full(l.shape, 1 / 9, np.float32) for l in backend._levels])
            backend.step(1)
            self.assertEqual(len(accumulations(m)), 2)
@@ -109,18 +109,17 @@ class ExchangeWindowGPU(unittest.TestCase):
        from test_raw_exchange_gpu import BOUNDS as RB, CFG, circle as rc, jet, state
        from CelerisLab.cuda.context import CudaContext
        with CudaContext(0) as ctx:
-            kw = dict(context=ctx, refinements=[RB], level_blocks=[(12, 8), (16, 20)],
-                      level_objects=[(), (rc(),)], level_jets=[(), (jet(),)],
+            kw = dict(context=ctx, fine_block=(16, 20), fine_objects=[rc()], jets=[jet()],
                      observe_raw_exchange=True)
-            with FixedPatchStepper(CFG, observe_window_exchange=True, **kw) as windowed, \
-                 FixedPatchStepper(CFG, **kw) as plain:
+            with FixedPatchStepper(CFG, RB, observe_window_exchange=True, **kw) as windowed, \
+                 FixedPatchStepper(CFG, RB, **kw) as plain:
                arrays = [state(level.shape) for level in windowed._levels]
                windowed.initialize(*arrays)
                plain.initialize(*arrays)
                windowed.step(2)
                plain.step(2)
                windowed.start_exchange_window()
-                b = windowed._device_boundaries[1]
+                b = windowed._level_boundaries[1]
                count = len(b.tables["cell"])
                factors = np.array([.25, .25, .125, .25])  # h*dt, h*h*dt, h*h for level-1 h=.5
                scale = np.zeros(4)
@@ -201,18 +200,18 @@ class ExchangeWindowShiftedMRTGPU(unittest.TestCase):
        from CelerisLab.cuda.context import CudaContext
        cfg = LBMConfig(nx=32, ny=24, viscosity=.1, collision="MRT", ddf_shifting=True)
        with CudaContext(0) as ctx:
-            kw = dict(context=ctx, refinements=[RB], level_blocks=[(12, 8), (16, 20)],
-                      level_objects=[(), (rc(),)], observe_raw_exchange=True)
-            with FixedPatchStepper(cfg, observe_window_exchange=True, **kw) as windowed, \
-                 FixedPatchStepper(cfg, observe_window_exchange=True, **kw) as split, \
-                 FixedPatchStepper(cfg, **kw) as plain:
+            kw = dict(context=ctx, base_block=(12, 8), fine_block=(16, 20), fine_objects=[rc()],
+                      observe_raw_exchange=True)
+            with FixedPatchStepper(cfg, RB, observe_window_exchange=True, **kw) as windowed, \
+                 FixedPatchStepper(cfg, RB, observe_window_exchange=True, **kw) as split, \
+                 FixedPatchStepper(cfg, RB, **kw) as plain:
                arrays = [state(level.shape) for level in windowed._levels]
                for run in (windowed, split, plain):
                    run.initialize(*arrays)
                    run.step(2)
                    if run is not plain:
                        run.start_exchange_window()
-                b = windowed._device_boundaries[1]
+                b = windowed._level_boundaries[1]
                count = len(b.tables["cell"])
                record = dict(scale=np.zeros(4), substeps=0)
                audit_substeps(windowed, b, count, record)
@@ -228,7 +227,7 @@ class ExchangeWindowShiftedMRTGPU(unittest.TestCase):
                    same_bits(self, a, e)
                rows = [np.empty((count, 4), np.float64) for _ in range(2)]
                cuda.memcpy_dtoh(rows[0], b.window_integral)
-                cuda.memcpy_dtoh(rows[1], split._device_boundaries[1].window_integral)
+                cuda.memcpy_dtoh(rows[1], split._level_boundaries[1].window_integral)
                np.testing.assert_array_equal(rows[0].view(np.uint64), rows[1].view(np.uint64))

                out = windowed.read_exchange_window(0)
@@ -1,17 +1,13 @@
-"""Two coarse/four fine steps: isolated fine circle on blocked storage, no flux-accuracy claim.
+"""Two coarse/four fine steps: isolated fine circle, no flux-accuracy claim.

-Run only through the bounded GPU runner after a fresh resource check. The fine
-step hook reads the actual output buffer; it never changes the stepping order.
-
-Retired with the dense refinement route (accepted losses, see the D2 commit):
-the dense-kernel per-hop flux ledger (fine/coarse transfer + boundary budget),
-the CurvedLinkSoA reflection oracle and the CELERIS_REFINEMENT_AUDIT collision
-counts. The blocked test keeps the four-hop comparison with the uniform solver,
-the raw-exchange mass sign and passive == zero actuation; raw exchange covers
-the boundary part only, not full population flux closure.
+Run only through the bounded GPU runner after a fresh resource check. The test
+hooks read actual kernel inputs/outputs; they never change the stepping order.
 """
 import importlib.util
+from contextlib import nullcontext
 from pathlib import Path
+from threading import RLock
+from types import SimpleNamespace
 from unittest.mock import Mock, patch

 import numpy as np
@@ -22,6 +18,7 @@ import CelerisLab
 from CelerisLab.body.geometry.circle import CircleGeometry
 from CelerisLab.body.objects import SimObject
 from CelerisLab.config import LBMConfig
+from CelerisLab.cuda import compiler_v2 as compiler
 from CelerisLab.cuda.context import CudaContext
 from CelerisLab.lbm import refinement_stepper as implementation
 from CelerisLab.lbm.descriptors import FLUID, SOLID
@@ -30,8 +27,10 @@ from CelerisLab.lbm.descriptors import FLUID, SOLID
 ROOT = Path(__file__).resolve().parents[3]
 CX = np.array([0, 1, -1, 0, 0, 1, -1, 1, -1])
 CY = np.array([0, 0, 0, 1, -1, 1, -1, -1, 1])
+OPP = np.array([0, 2, 1, 4, 3, 6, 5, 8, 7])
 W = np.array([4/9]+[1/9]*4+[1/36]*4)
 MOMENTS = np.stack((np.ones(9), CX, CY))
+GEOMETRY = ("fluid_idx", "dir", "q", "body_id", "rx", "ry", "rz", "fallback_class")
 BOUNDS = (4, 4, 20, 20)


@@ -54,6 +53,196 @@ def object_spec(speed):
                     radius=2.15, surface_actuation=surface(speed))


+def audit_compiler(original):
+    def compile_private(*args, **kwargs):
+        entry = Path(kwargs["build_dir"])/"fine_surface_audit.cu"
+        entry.write_text('#define CELERIS_REFINEMENT_AUDIT\n#include "refinement.cu"\n')
+        kwargs["source"] = entry.name
+        return original(*args, **kwargs)
+    return compile_private
+
+
+def reflection_oracle(before, cl, speed, *, verify_projection=True):
+    """FP64 BFL from frozen spatial populations, independent arc projection."""
+    flat = before.astype(float).reshape(9, -1)
+    radius = np.hypot(cl.rx.astype(float), cl.ry.astype(float))
+    delta = (np.arctan2(cl.ry.astype(float), cl.rx.astype(float))-.7+np.pi) % (2*np.pi)-np.pi
+    amplitude = 0 if speed is None else speed
+    strength = amplitude*np.maximum(0, 1-(delta/.85)**2)**2
+    velocity = strength[:, None]*np.column_stack((cl.rx, cl.ry))/radius[:, None]
+    if verify_projection and cl.surface_velocity is not None:
+        np.testing.assert_allclose(cl.surface_velocity, velocity, atol=3e-9, rtol=0)
+    incoming = []
+    for i, (k, d, q, fallback) in enumerate(zip(cl.fluid_idx, cl.dir, cl.q, cl.fallback_class)):
+        k, d, q = int(k), int(d), float(q)
+        rho = flat[:, k].sum()
+        dot = CX[d]*velocity[i, 0]+CY[d]*velocity[i, 1]
+        if fallback:
+            value = flat[d, k]-6*W[d]*rho*dot
+        elif q < .5:
+            value = 2*q*flat[d, k]+(1-2*q)*flat[d, k-CX[d]-40*CY[d]]-6*W[d]*rho*dot
+        else:
+            value = flat[d, k]/(2*q)+(1-1/(2*q))*flat[OPP[d], k]-3*W[d]*rho*dot/q
+        incoming.append(value)
+    return np.array(incoming)
+
+
+def fine_flux(before, staged, active, solid):
+    """Population ledger across the fine leaf boundary: solid versus halo.
+
+    Internal fluid-to-fluid streaming cancels pairwise; solid and halo packets
+    are counted in their actual direction, before any collision.
+    """
+    boundary, transfer = np.zeros(9), np.zeros(9)
+    magnitude = 0.0
+    for d, (cx, cy) in enumerate(zip(CX, CY)):
+        incoming = active & ~np.roll(active, (cy, cx), (0, 1))
+        outgoing = active & ~np.roll(active, (-cy, -cx), (0, 1))
+        from_solid = np.roll(solid, (cy, cx), (0, 1))
+        into_solid = np.roll(solid, (-cy, -cx), (0, 1))
+        received = np.roll(staged[d], (cy, cx), (0, 1)).astype(float)
+        sent = before[d].astype(float)
+        for vector, inc, out in ((boundary, incoming & from_solid, outgoing & into_solid),
+                                 (transfer, incoming & ~from_solid, outgoing & ~into_solid)):
+            assert np.isfinite(received[inc]).all() and np.isfinite(sent[out]).all()
+            vector[d] = received[inc].sum()-sent[out].sum()
+            magnitude += np.abs(received[inc]).sum()+np.abs(sent[out]).sum()
+    return boundary, transfer, magnitude
+
+
+def coarse_flux(before, fine, active):
+    """Derive every replacement from physical midpoints, not production tables."""
+    x0, y0, x1, y1 = BOUNDS
+    incoming = np.stack([np.roll(before[d], (CY[d], CX[d]), (0, 1)) for d in range(9)]).astype(float)
+    magnitude = 0.0
+    for y, x in np.argwhere(active):
+        for d in range(1, 9):
+            if x0 <= x+.5-CX[d]/2 <= x1 and y0 <= y+.5-CY[d]/2 <= y1:
+                a, b = 4+2*(x-x0), 4+2*(y-y0)
+                children = fine[d, b:b+2, a:a+2].astype(float)
+                assert children.shape == (2, 2) and np.isfinite(children).all()
+                incoming[d, y, x] = children.sum()/4
+                magnitude += np.abs(children).sum()/4
+    assert np.isfinite(incoming[:, active]).all()
+    # This retains the eight grazing-corner replacements of uncovered donors.
+    flux = (incoming[:, active]-before[:, active].astype(float)).sum(axis=1)
+    magnitude += np.abs(incoming[:, active]).sum()+np.abs(before[:, active]).sum(dtype=float)
+    return flux, magnitude
+
+
+def run_patch(speed, *, audit=True):
+    cfg = LBMConfig(nx=24, ny=24, viscosity=.05, velocity=0, rho=1)
+    coarse_active = np.ones((24, 24), bool)
+    coarse_active[4:20, 4:20] = False
+    source = object_spec(speed)
+    state = vars(source.state).copy()
+    with CudaContext(0) as context:
+        build = (patch.object(compiler, "compile_kernel", audit_compiler(compiler.compile_kernel))
+                 if audit else nullcontext())
+        with build:
+            backend = implementation.FixedPatchStepper(cfg, BOUNDS, context=context, fine_objects=[source])
+        with backend, context.activate():
+            c, f = backend._levels
+            cl = backend._fine_curved
+            geometry = {name: getattr(cl, name).copy() for name in GEOMETRY}
+            flags = backend.fine_boundary.flags.reshape(40, 40)
+            solid = (flags & SOLID) != 0
+            active = np.zeros((40, 40), bool)
+            active[4:-4, 4:-4] = True
+            active &= (flags & FLUID) != 0
+            active &= ~solid
+            local = backend.fine_boundary.objects[0]
+            assert local.center == (20., 20.) and local.radius == 4.3
+            assert source.center == (12.25, 12.25) and source.radius == 2.15
+            yy, xx = np.mgrid[:40, :40]
+            roi = active & ((xx-20)**2+(yy-20)**2 <= 9**2)
+            backend.initialize(rest((24, 24)), rest((40, 40)))
+            if audit:
+                for level in backend._levels:
+                    for name in ("collision_counts", "explosion_counts", "incoming_counts"):
+                        pointer, size = level.module.get_global(name)
+                        cuda.memset_d8(pointer, 0, size)
+            initial = backend.download()
+            initial_pop = initial[0][:, coarse_active].sum(axis=1, dtype=float)
+            initial_pop += .25*initial[1][:, active].sum(axis=1, dtype=float)
+            boundary, fine_transfer, coarse_transfer = np.zeros(9), np.zeros(9), np.zeros(9)
+            budget_sum, reflection_error, drive_mass = 0., 0., 0.
+            traces = []
+            original_curved, original_fine, original_coarse = backend._curved, f.kernel, c.kernel
+
+            def curved(*args, **kwargs):
+                nonlocal boundary, fine_transfer, budget_sum, reflection_error, drive_mass
+                backend._stream.synchronize()
+                before = read(args[0], f.shape)
+                expected = reflection_oracle(before, cl, speed)
+                original_curved(*args, **kwargs)
+                backend._stream.synchronize()
+                staged = read(args[0], f.shape)
+                d, k = cl.dir.astype(int), cl.fluid_idx.astype(int)
+                values = staged.reshape(9, -1)[OPP[d], k+CX[d]+40*CY[d]]
+                passive = reflection_oracle(before, cl, None, verify_projection=False)
+                drive_mass += .25*float(np.sum(values-passive))
+                error = float(np.max(np.abs(values-expected)))
+                reflection_error = max(reflection_error, error)
+                assert error <= 1e-7
+                np.testing.assert_array_equal(staged[:, active], before[:, active])
+                b, transfer, magnitude = fine_flux(before, staged, active, solid)
+                boundary += .25*b
+                fine_transfer += .25*transfer
+                budget_sum += .25*magnitude
+
+            def fine(*args, **kwargs):
+                nonlocal budget_sum
+                backend._stream.synchronize()
+                staged = read(args[0], f.shape)
+                original_fine(*args, **kwargs)
+                backend._stream.synchronize()
+                advanced = read(args[1], f.shape)
+                assert np.isfinite(advanced[:, active]).all()
+                assert np.isnan(advanced[:, solid]).all()
+                # Include actual collision operands in the FP32 rounding budget.
+                budget_sum += .25*np.abs(staged[:, active]).sum(dtype=float)
+                traces.append(advanced[:, roi].copy())
+
+            def coarse(*args, **kwargs):
+                nonlocal coarse_transfer, budget_sum
+                backend._stream.synchronize()
+                before, fine_values = read(args[0], c.shape), read(args[2], f.shape)
+                transfer, magnitude = coarse_flux(before, fine_values, coarse_active)
+                coarse_transfer += transfer
+                budget_sum += magnitude
+                return original_coarse(*args, **kwargs)
+
+            with patch.object(backend, "_curved", curved), patch.object(f, "kernel", fine), \
+                    patch.object(c, "kernel", coarse):
+                backend.step(2)
+            actual = backend.download()
+            population = actual[0][:, coarse_active].sum(axis=1, dtype=float)
+            population += .25*actual[1][:, active].sum(axis=1, dtype=float)
+            ledger = boundary+fine_transfer+coarse_transfer
+            error = MOMENTS@(population-initial_pop-ledger)
+            budget = 8*np.finfo(np.float32).eps*budget_sum
+            assert np.isfinite(budget_sum) and budget_sum > 0
+            assert np.max(np.abs(error)) <= budget
+            if speed:
+                assert speed*drive_mass > 0
+            assert len(traces) == 4 and backend.step_count == 2
+            if audit:
+                for level, mask, count in ((c, coarse_active, 2), (f, active, 4)):
+                    pointer, size = level.module.get_global("collision_counts")
+                    counts = read(pointer, mask.shape, np.uint32)
+                    assert size == counts.nbytes
+                    np.testing.assert_array_equal(counts, count*mask)
+            for name, values in geometry.items():
+                np.testing.assert_array_equal(getattr(cl, name), values)
+            assert vars(source.state) == state and source.geometry.center == source.center
+            print(f"speed={speed} audit={audit} BFL={reflection_error:.9g} ledger={error.tolist()} "
+                  f"budget={budget:.9g} boundary={list(MOMENTS@boundary)} "
+                  f"transfer={list(MOMENTS@(fine_transfer+coarse_transfer))} "
+                  f"drive_mass={drive_mass:+.9g}; counts={'PASS' if audit else 'normal build'}", flush=True)
+            return traces, roi, actual
+
+
 def uniform_reference(speed, roi):
    # Reuse the existing explicit config helper without importing unrelated tests.
    spec = importlib.util.spec_from_file_location("fine_surface_cases", ROOT/"tests/integration/_cases.py")
@@ -71,59 +260,61 @@ def uniform_reference(speed, roi):
        return traces


-def run_blocked_patch(speed):
-    """All-blocked run (whole-level tiles) of the same case; fine ROI after each fine hop."""
-    cfg = LBMConfig(nx=24, ny=24, viscosity=.05, velocity=0, rho=1)
-    source = object_spec(speed)
-    with CudaContext(0) as context:
-        with implementation.FixedPatchStepper(cfg, context=context, refinements=[BOUNDS],
-                                              level_blocks=[(24, 24), (40, 40)],
-                                              level_objects=[(), (source,)],
-                                              observe_raw_exchange=True) as backend:
-            c, f = backend._levels
-            flags = backend.fine_boundary.flags.reshape(40, 40)
-            active = np.zeros((40, 40), bool)
-            active[4:-4, 4:-4] = True
-            active &= ((flags & FLUID) != 0) & ((flags & SOLID) == 0)
-            yy, xx = np.mgrid[:40, :40]
-            roi = active & ((xx-20)**2+(yy-20)**2 <= 9**2)
-            backend.initialize(rest((24, 24)), rest((40, 40)))
-            traces, original = [], f.kernel
-
-            def fine(*args, **kwargs):
-                original(*args, **kwargs)
-                backend._stream.synchronize()
-                advanced = f.layout.to_dense(read(args[1], f.layout.blocked_shape))
-                assert np.isfinite(advanced[:, active]).all()
-                assert np.isnan(advanced[:, (flags & SOLID) != 0]).all()
-                traces.append(advanced[:, roi].copy())
-
-            with patch.object(f, "kernel", fine):
-                backend.step(2)
-            assert len(traces) == 4 and backend.step_count == 2
-            mass = backend.read_raw_exchange(0)["total"]["mass_into_fluid"]
-            return traces, roi, backend.download(), float(mass)
-
-
@pytest.mark.parametrize("speed", [0., .002, -.002])
-def test_fine_surface_blocked_matches_uniform_four_hops(speed):
-    """Blocked storage: fine ROI after each of four hops against the uniform 48^2 solver
-    (same tolerances as the dense test); blowing/suction drives mass with its sign;
-    zero actuation equals the passive circle bit for bit."""
-    actual, roi, final, mass = run_blocked_patch(speed)
+def test_fine_surface_four_hops(speed):
+    assert Path(CelerisLab.__file__).resolve() == ROOT/"src/CelerisLab/__init__.py"
+    assert Path(compiler._KERNEL_DIR).resolve() == ROOT/"src/CelerisLab/lbm/kernels"
+    assert Path(implementation.__file__).resolve() == ROOT/"src/CelerisLab/lbm/refinement_stepper.py"
+    actual, roi, final = run_patch(speed)
    reference = uniform_reference(speed, roi)
    for got, wanted in zip(actual, reference):
        np.testing.assert_allclose(got, wanted, atol=3e-6, rtol=0)
        np.testing.assert_allclose(MOMENTS@got.astype(float), MOMENTS@wanted.astype(float),
                                   atol=2e-5, rtol=0)
-    if speed:
-        assert speed*mass > 0
-    else:
-        passive, passive_roi, passive_final, _ = run_blocked_patch(None)
+    if speed == 0:
+        passive, passive_roi, passive_final = run_patch(None, audit=False)
        np.testing.assert_array_equal(passive_roi, roi)
        for got, wanted in zip(actual+list(final), passive+list(passive_final)):
            np.testing.assert_array_equal(got, wanted)
-    print(f"blocked speed={speed} four hops vs uniform PASS mass_into_fluid={mass:+.9g}", flush=True)
+
+
+@pytest.mark.parametrize("missing", ["surface_velocity_gpu", "link_exchange_gpu"])
+def test_missing_surface_device_buffer_rejects_before_launch(missing):
+    backend = implementation.FixedPatchStepper.__new__(implementation.FixedPatchStepper)
+    backend.fine_boundary = SimpleNamespace(objects=[object_spec(.002)])
+    backend._fine_curved = SimpleNamespace(fluid_idx_gpu=object(),
+                                          surface_velocity_gpu=object(), link_exchange_gpu=object())
+    setattr(backend._fine_curved, missing, None)
+    backend._curved = Mock()
+    with pytest.raises(RuntimeError, match="Surface boundary device buffers are missing"):
+        backend._apply_fine_boundary(object())
+    backend._curved.assert_not_called()
+
+
+def test_close_frees_fine_boundary_after_stream_failure_once():
+    backend = implementation.FixedPatchStepper.__new__(implementation.FixedPatchStepper)
+    backend.ctx = SimpleNamespace(activate=nullcontext)
+    backend._lock = RLock()
+    backend._status = "ready"
+    stream, soa, first, second = Mock(), Mock(), Mock(), Mock()
+    stream.synchronize.side_effect = RuntimeError("stream synchronization failed")
+    backend._stream = stream
+    backend._fine_curved = soa
+    backend._fine_flags = first
+    backend._curved = Mock()
+    backend._levels = [object(), object()]
+    backend._allocations = [first, second]
+    with pytest.raises(RuntimeError, match="stream synchronization failed"):
+        backend.close()
+    assert backend._status == "closed"
+    assert backend._levels == [] and backend._allocations == []
+    assert backend._fine_curved is None and backend._fine_flags is None
+    assert backend._curved is None and backend._stream is None
+    backend.close()
+    stream.synchronize.assert_called_once_with()
+    soa.free.assert_called_once_with()
+    first.free.assert_called_once_with()
+    second.free.assert_called_once_with()


 def test_cross_interface_donor_rejects_before_cuda():
@@ -134,8 +325,6 @@ def test_cross_interface_donor_rejects_before_cuda():
    context.activate.side_effect = AssertionError("must reject before CUDA activation")
    with patch.object(cuda, "mem_alloc") as allocate:
        with pytest.raises(ValueError, match="donor crosses"):
-            implementation.FixedPatchStepper(cfg, context=context, refinements=[BOUNDS],
-                                             level_blocks=[(24, 24), (40, 40)],
-                                             level_objects=[(), (source,)])
+            implementation.FixedPatchStepper(cfg, BOUNDS, context=context, fine_objects=[source])
        context.activate.assert_not_called()
        allocate.assert_not_called()
@@ -48,8 +48,9 @@ def make(context):
        for (x, y), amplitude in zip(centers, (.003, -.002))
    )
    run = FixedPatchStepper(
-        CFG, context=context, refinements=[PARENTS, START], level_blocks=[(4, 4)] * 3,
-        level_objects=[(), (), objects], level_jets=[(), (), jets],
+        CFG, PARENTS, context=context, base_block=(4, 4), fine_block=(4, 4),
+        child_block=(4, 4), child_rectangles_in_parent=START,
+        child_objects=objects, child_jets=jets,
    )
    return run, objects

@@ -25,8 +25,9 @@ B = ((16,16,32,40), (60,16,76,40))
 def make(ctx):
    body = SimObject(71, geometry=CircleGeometry(12.,14.,.8),center=(12.,14.),radius=.8)
    jet = Jet(Tube((10.5,14.),0.,3.,.3),Parallel(0.),'parabolic',.003)
-    run = FixedPatchStepper(CFG,context=ctx,refinements=[PARENTS,A],level_blocks=[(4,4)]*3,
-                           level_objects=[(),(),(body,)],level_jets=[(),(),(jet,)])
+    run = FixedPatchStepper(CFG,PARENTS,context=ctx,base_block=(4,4),fine_block=(4,4),
+                           child_block=(4,4),child_rectangles_in_parent=A,
+                           child_objects=(body,),child_jets=(jet,))
    return run,(body,)


@@ -90,22 +90,18 @@ def census(run, index, buffer):
    return totals(dense, leaf), external, independent


-
-def fine(block):
-    """Two-level circle+jet construction on BOUNDS, blocked with a (12, 8) root tile."""
-    return dict(refinements=[BOUNDS], level_blocks=[(12, 8), block],
-                level_objects=[(), (circle(),)], level_jets=[(), (jet(),)])
-
 class RawExchangeGPU(unittest.TestCase):
    def test_row_ledger_interval_jet_switch_and_no_population_effect(self):
        with CudaContext(0) as ctx:
            for block in ((32, 40), (16, 20), (8, 10)):
-                with FixedPatchStepper(CFG, context=ctx, **fine(block), observe_raw_exchange=True) as measured, \
-                     FixedPatchStepper(CFG, context=ctx, **fine(block)) as plain:
+                with FixedPatchStepper(CFG, BOUNDS, context=ctx, fine_block=block,
+                        fine_objects=[circle()], jets=[jet()], observe_raw_exchange=True) as measured, \
+                     FixedPatchStepper(CFG, BOUNDS, context=ctx, fine_block=block,
+                        fine_objects=[circle()], jets=[jet()]) as plain:
                    arrays = [state(l.shape) for l in measured._levels]
                    measured.initialize(*arrays)
                    plain.initialize(*arrays)
-                    b = measured._device_boundaries[1]
+                    b = measured._level_boundaries[1]
                    count = len(b.tables['cell'])
                    kernel = measured._levels[1].kernel
                    samples = []
@@ -146,6 +142,27 @@ class RawExchangeGPU(unittest.TestCase):
                        np.testing.assert_array_equal(measured.read_raw_exchange(0)['total']['impulse'],
                                                      obs['total']['impulse'])

+    def test_dense_fine_storage_reference(self):
+        """Same fine resolution/initial data with dense storage: a numerical
+        reference for cut-link rows, not a uniform-whole-domain convergence study."""
+        with CudaContext(0) as ctx:
+            with FixedPatchStepper(CFG, BOUNDS, context=ctx, fine_objects=[circle()], jets=[jet()]) as dense, \
+                 FixedPatchStepper(CFG, BOUNDS, context=ctx, fine_block=(16, 20),
+                    fine_objects=[circle()], jets=[jet()], observe_raw_exchange=True) as blocked:
+                arrays = [state(l.shape) for l in blocked._levels]
+                dense.initialize(*arrays)
+                blocked.initialize(*arrays)
+                for speed in (0., .002, 0.):
+                    dense.set_jet_amplitude(0, speed)
+                    blocked.set_jet_amplitude(0, speed)
+                    dense.step(); blocked.step()
+                    for a, e in zip(blocked.download(), dense.download()):
+                        np.testing.assert_array_equal(a, e)
+                    count = dense._fine_curved.count
+                    expected = download(dense._fine_curved.link_exchange_gpu, (count, 3))
+                    actual = download(blocked._level_boundaries[1].last_raw, (count, 4))
+                    np.testing.assert_array_equal(actual[:, :3], expected)
+
    def test_depth_four_interval_storage_and_sampling(self):
        cfg = LBMConfig(nx=48, ny=48, viscosity=.02)
        obj = SimObject(0, geometry=CircleGeometry(16., 16., .5), center=(16., 16.), radius=.5)
@@ -158,7 +175,7 @@ class RawExchangeGPU(unittest.TestCase):
                 FixedPatchStepper(cfg, context=ctx, **settings) as plain:
                arrays = [state(l.shape, ux=0.) for l in run._levels]
                run.initialize(*arrays); plain.initialize(*arrays)
-                b = run._device_boundaries[3]
+                b = run._level_boundaries[3]
                count = len(b.tables['cell'])
                kernel = run._levels[3].kernel
                samples, independent_samples = [], []
@@ -204,35 +221,35 @@ class RawExchangeGPU(unittest.TestCase):
        Eight local hops are a short-time equivalence check, not convergence.
        The test adapter only enables the standalone kernel's optional inputs.
        """
+        from CelerisLab.lbm.blocks import BlockLayout2D
+        from CelerisLab.lbm.block_stepper import PeriodicBlockStepper
        fine_cfg = replace(CFG, nx=64, ny=48, viscosity=.2)
        obj = SimObject(0, geometry=CircleGeometry(31.5, 23.5, 4.3),
                        center=(31.5, 23.5), radius=4.3)
+        layout = BlockLayout2D.build(48, 64, 8, 8)
        with CudaContext(0) as ctx:
-            with FixedPatchStepper(fine_cfg, context=ctx, refinements=[], level_blocks=[(8, 8)],
-                                   level_objects=[(obj,)]) as uniform, \
-                 FixedPatchStepper(CFG, context=ctx, **fine((16, 20)), observe_raw_exchange=True) as amr:
+            with PeriodicBlockStepper(fine_cfg, layout, context=ctx, objects=[obj]) as uniform, \
+                 FixedPatchStepper(CFG, BOUNDS, context=ctx, fine_block=(16, 20),
+                    fine_objects=[circle()], jets=[jet()], observe_raw_exchange=True) as amr:
                amr.initialize(*(state(l.shape) for l in amr._levels))
                uniform.initialize(state((9, 48, 64)))
-                b = amr._device_boundaries[1]
+                b = amr._level_boundaries[1]
                count = len(b.tables['cell'])
                links = obj.geometry.build_cut_links(64, 48)
                np.testing.assert_array_equal([v.dir for v in links], b.tables['dir'])
                np.testing.assert_allclose([v.q for v in links], b.tables['q'], atol=2e-7, rtol=0)
-                # Same fine-grid force arms after translation to the same center: the
-                # single level's static BFL gets the AMR Jet column and force arms; the raw
-                # output pointer is substituted at launch (no stepper-side accumulation).
-                root = uniform._device_boundaries[0]
+                # Same fine-grid force arms after translation to the same center.
+                rx = uniform._upload(b.tables['rx'])
+                ry = uniform._upload(b.tables['ry'])
                velocity = uniform._upload(b.tables['surface_velocity'])
-                root.device.update(surface_velocity=velocity, rx=uniform._upload(b.tables['rx']),
-                                   ry=uniform._upload(b.tables['ry']))
                raw = uniform._allocate(count*16)
-                original = root.kernel
+                original = uniform._curved
                samples = []
                def measured_boundary(*args, **kw):
-                    original(*args[:10], raw, **kw)  # BlockCurvedBoundary raw-exchange output
+                    original(*args[:7], velocity, rx, ry, raw, **kw)
                    uniform._stream.synchronize()
                    samples.append(download(raw, (count, 4)))
-                root.kernel = measured_boundary
+                uniform._curved = measured_boundary
                for speed in (0., .002, -.001, 0.):
                    amr.set_jet_amplitude(0, speed)
                    cuda.memcpy_htod(velocity, b.tables['surface_velocity'])
@@ -2,12 +2,6 @@

 Run CPU and GPU classes in separate processes. GPU callers must first check
 resources/device availability; this file never selects or reserves a GPU.
-
-Dense refinement storage was removed: every level is blocked and level_blocks
-is required (TileContractCPU). The retired dense/blocked route parity gates
-(BlockedBaseGPU exact parity, DenseBlockedRouteParity incl. MRT and two
-bodies, the route capability matrix) passed and are recorded at 1573942 and
-2469af7.
 """
 from contextlib import ExitStack, contextmanager, nullcontext
 from dataclasses import replace
@@ -19,13 +13,12 @@ from unittest.mock import Mock, patch
 import numpy as np
 import pycuda.driver as cuda

-from CelerisLab.cuda import compiler
+from CelerisLab.cuda import compiler_v2 as compiler
 from CelerisLab.lbm import refinement_stepper
 from CelerisLab.lbm.blocks import BlockLayout2D
 from CelerisLab.lbm.descriptors import D2Q9_EX, D2Q9_EY
 from CelerisLab.lbm.refinement import ACTIVE, OUTER_GHOST, FixedPatch2D
-from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
-from CelerisLab.lbm.refinement_tables import blocked_coarse_tables
+from CelerisLab.lbm.refinement_stepper import FixedPatchStepper, blocked_base_tables
 from test_refinement_blocked_fine import NX, NY, BOUNDS, config, state, circle, actuation

 BASE_BLOCKS = ((24, 32), (12, 8), (3, 4), (1, 1))
@@ -80,7 +73,6 @@ def mocked_backend(mode="periodic", fail_alloc=None):
            stack.enter_context(patch.object(compiler, name))
        stack.enter_context(patch.object(compiler, "load_module", return_value=module))
        stack.enter_context(patch.object(cuda, "Stream", return_value=stream))
-        stack.enter_context(patch.object(cuda, "mem_get_info", return_value=(1 << 40, 1 << 40)))
        stack.enter_context(patch.object(cuda, "mem_alloc", side_effect=allocate))
        stack.enter_context(patch.object(cuda, "memcpy_htod",
                                        side_effect=lambda p, a: uploads.append((p, a))))
@@ -91,10 +83,11 @@ def mocked_backend(mode="periodic", fail_alloc=None):
 class BlockedBaseCPU(unittest.TestCase):
    def test_reject_before_cuda(self):
        context = Mock(activate=Mock(side_effect=AssertionError("CUDA reached")))
-        for blocks in ([(12, 8), None], [(7, 8), FINE_BLOCK], [(12,), FINE_BLOCK]):
-            with self.subTest(blocks=blocks), self.assertRaises(ValueError):
-                FixedPatchStepper(config("periodic"), context=context, refinements=[BOUNDS],
-                                  level_blocks=blocks)
+        for kwargs in ({"base_block": (12, 8)},
+                       {"base_block": (7, 8), "fine_block": FINE_BLOCK},
+                       {"base_block": (12,), "fine_block": FINE_BLOCK}):
+            with self.subTest(kwargs=kwargs), self.assertRaises(ValueError):
+                FixedPatchStepper(config("periodic"), BOUNDS, context=context, **kwargs)
        context.activate.assert_not_called()

    def test_all_arrival_rows_donors_and_kind_masks(self):
@@ -109,7 +102,7 @@ class BlockedBaseCPU(unittest.TestCase):
                leaves[[0, -1]] = False
            for by, bx in BASE_BLOCKS:
                layout = BlockLayout2D.build(NY, NX, by, bx, periodic=(True, mode == "periodic"))
-                donors, targets, kind = blocked_coarse_tables(p, layout, order, mode != "periodic")
+                donors, targets, kind = blocked_base_tables(p, layout, order, mode != "periodic")
                ly, lx = layout.local_shape
                L = ly * lx
                expected, source_cells = [], []
@@ -139,7 +132,7 @@ class BlockedBaseCPU(unittest.TestCase):
        p = FixedPatch2D.from_config(config("periodic"), BOUNDS)
        order = np.argsort(p.coalesce_coarse)
        layout = BlockLayout2D.build(NY, NX, 3, 4)
-        donors, targets, _ = blocked_coarse_tables(p, layout, order)
+        donors, targets, _ = blocked_base_tables(p, layout, order)
        dense = np.arange(9 * NY * NX, dtype=np.float32).reshape(9, NY, NX)
        initial = layout.to_blocked(dense)
        arrivals = np.arange(len(targets), dtype=np.float32) + 10000
@@ -179,7 +172,7 @@ class BlockedBaseCPU(unittest.TestCase):
        # With one parent block, all eight grazing sources are owner cores.
        # Each is read for four fine ghost children: late explosion corrupts32.
        single = BlockLayout2D.build(NY, NX, NY, NX)
-        donors, targets, _ = blocked_coarse_tables(p, single, order)
+        donors, targets, _ = blocked_base_tables(p, single, order)
        initial = single.to_blocked(dense)
        staged = initial.copy()
        staged.ravel()[targets.astype(np.intp)] = arrivals
@@ -200,10 +193,9 @@ class BlockedBaseCPU(unittest.TestCase):

    def test_launch_order_stream_wall_constants_and_buffer_swaps(self):
        with mocked_backend("slot_reservoir") as m:
-            backend = FixedPatchStepper(config("slot_reservoir"), context=m["context"],
-                                        coarse_boundary="slot_reservoir", refinements=[BOUNDS],
-                                        level_blocks=[(3, 4), FINE_BLOCK],
-                                        level_objects=[(), (circle(actuation(.002)),)])
+            backend = FixedPatchStepper(config("slot_reservoir"), BOUNDS, context=m["context"],
+                                        coarse_boundary="slot_reservoir", fine_block=FINE_BLOCK,
+                                        base_block=(3, 4), fine_objects=[circle(actuation(.002))])
            try:
                backend.initialize(state(NY, NX, 1), state(*backend.patch.fine_shape, 2))
                coarse, fine = backend._levels
@@ -231,8 +223,6 @@ class BlockedBaseCPU(unittest.TestCase):
                    self.assertIs(collide[4], coarse.kind)
                walls = [a for ptr, a in m["uploads"] if isinstance(ptr, str) and ptr == "d_block_wall"]
                self.assertEqual(len(walls), 1)
-                self.assertEqual([call.kwargs["block_root_bc"] for call in
-                                  compiler.generate_config.call_args_list], [1, 0])
                width, height, slot, speed = struct.unpack("=IIIf", walls[0])
                self.assertEqual((width, height, slot), (NX, NY, 8))
                self.assertAlmostEqual(speed, .001)
@@ -246,16 +236,16 @@ class BlockedBaseCPU(unittest.TestCase):
        for fail_at in (1, 6, 10):
            with self.subTest(fail_at=fail_at), mocked_backend(fail_alloc=fail_at) as m:
                with self.assertRaisesRegex(MemoryError, "injected allocation"):
-                    FixedPatchStepper(config("periodic"), context=m["context"], refinements=[BOUNDS],
-                                      level_blocks=[(3, 4), FINE_BLOCK])
+                    FixedPatchStepper(config("periodic"), BOUNDS, context=m["context"],
+                                      fine_block=FINE_BLOCK, base_block=(3, 4))
                self.assertEqual(len(m["allocations"]), fail_at)
                for allocation in m["allocations"]:
                    allocation.free.assert_called_once_with()

    def test_failed_parent_stage_cannot_publish_or_resume(self):
        with mocked_backend() as m:
-            backend = FixedPatchStepper(config("periodic"), context=m["context"], refinements=[BOUNDS],
-                                        level_blocks=[(3, 4), FINE_BLOCK])
+            backend = FixedPatchStepper(config("periodic"), BOUNDS, context=m["context"],
+                                        fine_block=FINE_BLOCK, base_block=(3, 4))
            backend.initialize(state(NY, NX, 1), state(*backend.patch.fine_shape, 2))
            m["functions"]["BlockCoalesce"].side_effect = RuntimeError("injected parent stage failure")
            current = backend._levels[0].current
@@ -273,33 +263,61 @@ class BlockedBaseCPU(unittest.TestCase):
                allocation.free.assert_called_once_with()


-class TileContractCPU(unittest.TestCase):
-    """level_blocks is required: one positive (by, bx) tile per level, checked before CUDA."""
+class BlockedBaseGPU(unittest.TestCase):
+    def test_bodyfree_exact_parity_and_raw_invalid_slots(self):
+        self.compare([(None, ((24, 32), (3, 4), (1, 1)), 8)])

-    def test_level_blocks_required_and_validated(self):
-        context = Mock(activate=Mock(side_effect=AssertionError("CUDA reached")))
-        cfg = config("periodic")
-        for label, kw in (("omitted", {}),
-                          ("None list", dict(level_blocks=None)),
-                          ("too few levels", dict(level_blocks=[(12, 8)])),
-                          ("too many levels", dict(level_blocks=[(12, 8), FINE_BLOCK, FINE_BLOCK])),
-                          ("None root", dict(level_blocks=[None, FINE_BLOCK])),
-                          ("None fine", dict(level_blocks=[(12, 8), None])),
-                          ("wrong arity", dict(level_blocks=[(12,), FINE_BLOCK])),
-                          ("non-integer", dict(level_blocks=[(12, 8.0), FINE_BLOCK])),
-                          ("bool", dict(level_blocks=[(True, 8), FINE_BLOCK])),
-                          ("non-positive", dict(level_blocks=[(0, 8), FINE_BLOCK])),
-                          ("non-tuple", dict(level_blocks=[12, FINE_BLOCK])),
-                          ("not dividing", dict(level_blocks=[(7, 8), FINE_BLOCK]))):
-            with self.subTest(label), self.assertRaises(ValueError):
-                FixedPatchStepper(cfg, context=context, refinements=[BOUNDS], **kw)
-        with self.assertRaisesRegex(ValueError, "whole-level tile"):
-            FixedPatchStepper(cfg, context=context, refinements=[BOUNDS])
-        context.activate.assert_not_called()
-        # The documented whole-level tiles are always valid (construction reaches CUDA).
-        whole = [(NY, NX), FixedPatch2D.from_config(cfg, BOUNDS).fine_shape]
-        with self.assertRaisesRegex(AssertionError, "CUDA reached"):
-            FixedPatchStepper(cfg, context=context, refinements=[BOUNDS], level_blocks=whole)
+    def test_actuated_circle_exact_parity_and_raw_invalid_slots(self):
+        self.compare([(.002, ((12, 8),), 8), (-.002, ((12, 8),), 8)])
+
+    def test_non_power_two_slot_width_exact_parity(self):
+        self.compare([(None, ((3, 4),), 7), (None, ((3, 4),), 10)], ("slot_reservoir",))
+
+    def compare(self, cases, modes=("periodic", "slot_reservoir")):
+        from CelerisLab.cuda.context import CudaContext
+        expected_source = Path(__file__).resolve().parents[3] / "src/CelerisLab/lbm/refinement_stepper.py"
+        self.assertEqual(Path(refinement_stepper.__file__).resolve(), expected_source)
+        with CudaContext(0) as context:
+            for mode in modes:
+                for speed, layouts, width in cases:
+                    cfg = config(mode)
+                    if mode == "slot_reservoir":
+                        cfg = replace(cfg, wall_slot_width=width)
+                    p = FixedPatch2D.from_config(cfg, BOUNDS)
+                    arrays = state(NY, NX, 1), state(*p.fine_shape, 2)
+                    reference = None
+                    for base_block in (None, *layouts):
+                        label = f"{mode} width={width} speed={speed} base={base_block}"
+                        snapshots = []
+                        with FixedPatchStepper(cfg, BOUNDS, context=context, coarse_boundary=mode,
+                                               fine_block=FINE_BLOCK, base_block=base_block,
+                                               fine_objects=() if speed is None else [circle(actuation(speed))]) as backend:
+                            backend.initialize(*arrays)
+                            for step in (0, 1, 2, 8):
+                                if step == 1 and base_block is not None:
+                                    # Blocked staging dirties invalid input slots; its output
+                                    # must poison them. Dense parent does not stage there.
+                                    coarse = backend._levels[0]
+                                    spare = coarse.layout.to_blocked(np.full(coarse.shape, 42, np.float32))
+                                    cuda.memcpy_htod(coarse.spare, spare)
+                                backend.step(step - backend.step_count)
+                                public, owner = backend.download(), raw(backend)
+                                self.assertEqual(backend.step_count, step)
+                                self.assertTrue(np.isnan(owner[0][:, 6:18, 8:24]).all(), label)
+                                if mode != "periodic":
+                                    self.assertTrue(np.isnan(owner[0][:, (0, -1), :]).all(), label)
+                                snapshots.append(public + owner)
+                            backend.initialize(*arrays)
+                            self.assertEqual(backend.step_count, 0)
+                            for actual, expected in zip(backend.download() + raw(backend), snapshots[0]):
+                                exact(self, actual, expected, label + " reset")
+                        if reference is None:
+                            reference = snapshots
+                        else:
+                            for step, actual, expected in zip((0, 1, 2, 8), snapshots, reference):
+                                for index, (a, b) in enumerate(zip(actual, expected)):
+                                    exact(self, a, b, f"{label} step={step} array={index}")
+                            print(f"PASS {label}: full finite bits/NaN masks at 0,1,2,8 + reset", flush=True)


 if __name__ == "__main__":
@@ -1,8 +1,9 @@
-"""Blocked fine storage for FixedPatchStepper: CPU tables, staging and launch contracts.
+"""Opt-in blocked fine storage for FixedPatchStepper: CPU tables plus GPU gate.

-The former GPU gate (dense fine storage vs 1x1, 2x2 and 4x2 fine blocks,
-bit-identical) was retired with the dense route; its result and the
-dense/blocked route parity are recorded at 1573942 and 2469af7.
+Frozen GPU gate (same stream, 8 coarse steps, nonequilibrium start), in both
+periodic and slot_reservoir modes: dense fine storage vs 1x1, 2x2 and 4x2
+fine blocks give bit-identical masked downloads AND bit-identical raw coarse
+and unpacked owner-core fine arrays, including patch-ghost values/NaN masks.
 """
 from contextlib import nullcontext
 import unittest
@@ -14,15 +15,14 @@ import pycuda.driver as cuda
 from CelerisLab.body.geometry.circle import CircleGeometry
 from CelerisLab.body.objects import SimObject, SurfaceActuation
 from CelerisLab.config import LBMConfig
-from CelerisLab.cuda import compiler
+from CelerisLab.cuda import compiler_v2 as compiler
 from CelerisLab.lbm import curved_links
 from CelerisLab.lbm.blocks import NO_DONOR, BlockLayout2D, blocked_link_tables
 from CelerisLab.lbm.descriptors import D2Q9_EX, D2Q9_EY
 from CelerisLab.lbm.descriptors import SOLID
 from CelerisLab.lbm.refinement import (ACTIVE, INNER_GHOST, OUTER_GHOST, FixedPatch2D,
                                       prepare_fine_boundary)
-from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
-from CelerisLab.lbm.refinement_tables import blocked_fine_tables
+from CelerisLab.lbm.refinement_stepper import FixedPatchStepper, blocked_fine_tables

 NX, NY, BOUNDS, STEPS = 32, 24, (8, 6, 24, 18), 8
 FINE_BLOCKS = ((32, 40), (16, 20), (16, 10))  # fine window 32x40: 1x1, 2x2, 4x2
@@ -88,13 +88,15 @@ def seam_cells(layout, mapped):


 def raw(backend):
-    """Unmasked device arrays of every level; blocked storage unpacked from owner cores."""
-    out = []
-    for level in backend._levels:
-        array = np.empty(level.shape if level.layout is None else level.layout.blocked_shape,
-                         np.float32)
-        cuda.memcpy_dtoh(array, level.current)
-        out.append(array if level.layout is None else level.layout.to_dense(array))
+    """Unmasked device arrays; blocked fine storage unpacked from owner cores."""
+    coarse, fine = backend._levels
+    out = [np.empty(coarse.shape, np.float32)]
+    out.append(np.empty(fine.shape if backend.fine_layout is None
+                        else backend.fine_layout.blocked_shape, np.float32))
+    cuda.memcpy_dtoh(out[0], coarse.current)
+    cuda.memcpy_dtoh(out[1], fine.current)
+    if backend.fine_layout is not None:
+        out[1] = backend.fine_layout.to_dense(out[1])
    return out


@@ -103,10 +105,9 @@ class BlockedFineCPU(unittest.TestCase):
        return Mock(activate=Mock(side_effect=AssertionError("CUDA reached")), sm_arch="sm_70")

    def test_rejections_before_cuda(self):
-        for kw in ({"level_blocks": [None, (7, 7)]}, {"level_blocks": [None, (16,)]}):
+        for kw in ({"fine_block": (7, 7)}, {"fine_block": (16,)}):
            with self.subTest(kw=kw), self.assertRaises(ValueError):
-                FixedPatchStepper(config("periodic"), context=self.forbidden_context(),
-                                  refinements=[BOUNDS], **kw)
+                FixedPatchStepper(config("periodic"), BOUNDS, context=self.forbidden_context(), **kw)

    def test_single_block_layout_allocates_no_empty_halo_tables(self):
        # A 1x1 non-periodic fine layout has zero halo pairs; CUDA rejects mem_alloc(0).
@@ -120,12 +121,12 @@ class BlockedFineCPU(unittest.TestCase):
                raise AssertionError("zero-byte device allocation")
            sizes.append(size)
            return Mock()
-        with patch.object(cuda, "Stream"), patch.object(cuda, "mem_get_info", return_value=(1 << 40, 1 << 40)), patch.object(compiler, "generate_config"), \
+        with patch.object(cuda, "Stream"), patch.object(compiler, "generate_config"), \
                patch.object(compiler, "compile_kernel"), patch.object(compiler, "cleanup_build_dir"), \
                patch.object(compiler, "load_module", return_value=module), \
                patch.object(cuda, "mem_alloc", side_effect=alloc), patch.object(cuda, "memcpy_htod"):
-            backend = FixedPatchStepper(config("periodic"), context=context, refinements=[BOUNDS],
-                                        level_blocks=[(12, 8), FINE_BLOCKS[0]])
+            backend = FixedPatchStepper(config("periodic"), BOUNDS, context=context,
+                                        fine_block=FINE_BLOCKS[0])
            self.assertEqual(len(backend.fine_layout.halo_dst), 0)
            self.assertIsNone(backend._levels[1].halo_dst)
            backend._gather_halo(backend._levels[1], Mock())  # no launch without pairs
@@ -170,15 +171,14 @@ class BlockedFineCPU(unittest.TestCase):
            return functions.setdefault(name, Mock(side_effect=lambda *a, **k: calls.append((name, a))))
        module = Mock(get_function=Mock(side_effect=get_function))
        module.get_global.return_value = (Mock(), 20)
-        with patch.object(cuda, "Stream"), patch.object(cuda, "mem_get_info", return_value=(1 << 40, 1 << 40)), patch.object(compiler, "generate_config"), \
+        with patch.object(cuda, "Stream"), patch.object(compiler, "generate_config"), \
                patch.object(compiler, "compile_kernel"), patch.object(compiler, "cleanup_build_dir"), \
                patch.object(compiler, "load_module", return_value=module), \
                patch.object(cuda, "mem_alloc", side_effect=lambda size: Mock()), \
                patch.object(cuda, "memcpy_htod"), \
                patch.object(curved_links, "CurvedLinkSoA", side_effect=AssertionError("dense SoA built")):
-            backend = FixedPatchStepper(config("periodic"), context=context, refinements=[BOUNDS],
-                                        level_blocks=[(12, 8), FINE_BLOCKS[1]],
-                                        level_objects=[(), tuple(objects)])
+            backend = FixedPatchStepper(config("periodic"), BOUNDS, context=context,
+                                        fine_objects=objects, fine_block=FINE_BLOCKS[1])
            backend.initialize(state(NY, NX, 1), state(*backend.patch.fine_shape, 2))
        return backend, calls

@@ -187,8 +187,8 @@ class BlockedFineCPU(unittest.TestCase):
            backend, calls = self.mocked_backend([circle(spec)])
            c, f = backend._levels
            coarse, cur, spare = c.current, f.current, f.spare
-            flags, kind = backend._device_boundaries[1].flags, backend._levels[1].kind
-            velocity = backend._device_boundaries[1].device.get("surface_velocity")
+            flags, kind = backend._fine_flags, backend._levels[1].kind
+            velocity = backend._link_tables.get("surface_velocity")
            self.assertEqual(velocity is None, spec is None)
            backend.step(1)
            backend.close()
@@ -196,9 +196,8 @@ class BlockedFineCPU(unittest.TestCase):
            self.assertEqual(order, [("BlockExplode", coarse),
                                     ("BlockHaloCopy", cur), ("BlockCurvedBoundary", cur), ("BlockStep", cur),
                                     ("BlockHaloCopy", spare), ("BlockCurvedBoundary", spare),
-                                     ("BlockStep", spare), ("BlockHaloCopy", coarse),
-                                     ("BlockCoalesce", cur), ("BlockStep", coarse)])
-            steps = [a for name, a in calls if name == "BlockStep"][:2]  # the two fine hops
+                                     ("BlockStep", spare), ("PatchCoarseStep", coarse)])
+            steps = [a for name, a in calls if name == "BlockStep"]
            self.assertEqual([a[1] for a in steps], [spare, cur])
            for a in steps:
                self.assertIs(a[3], flags)  # blocked flags, not null
@@ -214,7 +213,7 @@ class BlockedFineCPU(unittest.TestCase):
        for speed in (0.0, .002):  # any spec, even zero amplitude
            backend, calls = self.mocked_backend([circle(actuation(speed))])
            self.assertIsNotNone(backend._fine_tables["surface_velocity"])
-            del backend._device_boundaries[1].device["surface_velocity"]
+            del backend._link_tables["surface_velocity"]
            with self.assertRaisesRegex(RuntimeError, "Surface velocity device buffer is missing"):
                backend.step(1)
            names = [name for name, _ in calls]
@@ -280,5 +279,91 @@ class BlockedFineCPU(unittest.TestCase):
            self.assertEqual(set(np.unique(kind).tolist()) - {ACTIVE, INNER_GHOST, OUTER_GHOST}, set())


+class BlockedFineGPU(unittest.TestCase):
+    def test_dense_and_blocked_fine_storage_are_bit_identical(self):
+        from CelerisLab.cuda.context import CudaContext
+        for mode in ("periodic", "slot_reservoir"):
+            cfg = config(mode)
+            patch = FixedPatch2D.from_config(cfg, BOUNDS)
+            g_coarse, g_fine = state(NY, NX, 1), state(*patch.fine_shape, 2)
+            results = []
+            with CudaContext(0) as context:
+                for fine_block in (None, *FINE_BLOCKS):
+                    with FixedPatchStepper(cfg, BOUNDS, context=context, coarse_boundary=mode,
+                                           fine_block=fine_block) as backend:
+                        backend.initialize(g_coarse, g_fine)
+                        backend.step(STEPS)
+                        self.assertEqual(backend.step_count, STEPS)
+                        results.append((backend.download(), raw(backend)))
+            (dense_public, dense_raw), blocked = results[0], results[1:]
+            kind = patch.fine_kind
+            self.assertTrue(np.isnan(dense_raw[1][:, kind == OUTER_GHOST]).all(), mode)
+            # INNER populations may be NaN after hop 2 (pulled from poisoned OUTER);
+            # their exact pattern is pinned by the dense-vs-blocked comparison below.
+            self.assertTrue(np.isfinite(dense_raw[1][:, kind == ACTIVE]).all(), mode)
+            for fine_block, (public, raw_arrays) in zip(FINE_BLOCKS, blocked):
+                msg = f"{mode} fine_block={fine_block}"
+                for a, b in zip(public + tuple(raw_arrays), dense_public + tuple(dense_raw)):
+                    np.testing.assert_array_equal(a, b, err_msg=msg)  # values and NaN masks
+            print(f"{mode}: dense vs {len(blocked)} fine-block layouts bit-identical "
+                  f"after {STEPS} coarse steps", flush=True)
+
+    def test_passive_circle_dense_and_blocked_fine_storage_are_bit_identical(self):
+        from CelerisLab.cuda.context import CudaContext
+        for mode in ("periodic", "slot_reservoir"):
+            cfg = config(mode)
+            patch, boundary = fine_circle(mode)
+            solid = (boundary.flags.reshape(patch.fine_shape) & SOLID) != 0
+            g_coarse, g_fine = state(NY, NX, 1), state(*patch.fine_shape, 2)
+            results = []
+            with CudaContext(0) as context:
+                for fine_block in (None, *FINE_BLOCKS):
+                    with FixedPatchStepper(cfg, BOUNDS, context=context, coarse_boundary=mode,
+                                           fine_objects=[circle()], fine_block=fine_block) as backend:
+                        backend.initialize(g_coarse, g_fine)
+                        backend.step(STEPS)
+                        results.append((backend.download(), raw(backend)))
+            (dense_public, dense_raw), blocked = results[0], results[1:]
+            fine = dense_raw[1]
+            self.assertTrue(np.isfinite(fine[:, (patch.fine_kind == ACTIVE) & ~solid]).all(), mode)
+            self.assertTrue(np.isnan(fine[:, solid]).all(), mode)
+            self.assertTrue(np.isnan(fine[:, patch.fine_kind == OUTER_GHOST]).all(), mode)
+            for fine_block, (public, raw_arrays) in zip(FINE_BLOCKS, blocked):
+                msg = f"circle {mode} fine_block={fine_block}"
+                for a, b in zip(public + tuple(raw_arrays), dense_public + tuple(dense_raw)):
+                    np.testing.assert_array_equal(a, b, err_msg=msg)  # values and NaN masks
+            print(f"circle {mode}: dense vs {len(blocked)} fine-block layouts bit-identical "
+                  f"after {STEPS} coarse steps", flush=True)
+
+    def test_actuated_circle_dense_and_blocked_fine_storage_are_bit_identical(self):
+        from CelerisLab.cuda.context import CudaContext
+        for mode in ("periodic", "slot_reservoir"):
+            for speed in (.002, -.002):
+                cfg = config(mode)
+                patch, boundary = fine_circle(mode, actuation(speed))
+                solid = (boundary.flags.reshape(patch.fine_shape) & SOLID) != 0
+                g_coarse, g_fine = state(NY, NX, 1), state(*patch.fine_shape, 2)
+                results = []
+                with CudaContext(0) as context:
+                    for fine_block in (None, *FINE_BLOCKS):
+                        with FixedPatchStepper(cfg, BOUNDS, context=context, coarse_boundary=mode,
+                                               fine_objects=[circle(actuation(speed))],
+                                               fine_block=fine_block) as backend:
+                            backend.initialize(g_coarse, g_fine)
+                            backend.step(STEPS)
+                            results.append((backend.download(), raw(backend)))
+                (dense_public, dense_raw), blocked = results[0], results[1:]
+                fine = dense_raw[1]
+                self.assertTrue(np.isfinite(fine[:, (patch.fine_kind == ACTIVE) & ~solid]).all())
+                self.assertTrue(np.isnan(fine[:, solid]).all())
+                self.assertTrue(np.isnan(fine[:, patch.fine_kind == OUTER_GHOST]).all())
+                for fine_block, (public, raw_arrays) in zip(FINE_BLOCKS, blocked):
+                    msg = f"actuated {speed} {mode} fine_block={fine_block}"
+                    for a, b in zip(public + tuple(raw_arrays), dense_public + tuple(dense_raw)):
+                        np.testing.assert_array_equal(a, b, err_msg=msg)  # values and NaN masks
+                print(f"actuated {speed:+.3f} {mode}: dense vs {len(blocked)} fine-block layouts "
+                      f"bit-identical after {STEPS} coarse steps", flush=True)
+
+
 if __name__ == "__main__":
    unittest.main()
@@ -17,10 +17,9 @@ import pycuda.driver as cuda

 import CelerisLab
 from CelerisLab.lbm import refinement_stepper as implementation
-from CelerisLab.lbm.blocks import BlockLayout2D
 from test_refinement_gpu import (ROOT, POP_ATOL, MOMENT_ATOL, ref, postcollision,
                                 totals, LBMConfig, CudaContext, compiler,
-                                 FixedPatchStepper)
+                                 PeriodicPatchStepper)


 def scalar_interval(coarse, fine, bounds, slot=None):
@@ -84,21 +83,6 @@ def scalar_interval(coarse, fine, bounds, slot=None):
    return result, current


-def whole_tiles(cfg, bounds):
-    """One whole-level tile per level: (ny, nx) at the root, the fine window below."""
-    return [(cfg.ny, cfg.nx), implementation.FixedPatch2D.from_config(cfg, bounds).fine_shape]
-
-
-def owner_arrays(backend):
-    """Raw owner storage of every level as dense (9, ny, nx), before public masking."""
-    arrays = []
-    for level in backend._levels:
-        blocked = np.empty(level.layout.blocked_shape, np.float32)
-        cuda.memcpy_dtoh(blocked, level.current)
-        arrays.append(level.layout.to_dense(blocked))
-    return arrays
-
-
 class FaultGuardsCPU(unittest.TestCase):
    def test_boundary_mode_and_margin_contract(self):
        cfg = LBMConfig(nx=32, ny=24, viscosity=.1, velocity=0, wall_normal_velocity=.001,
@@ -110,8 +94,7 @@ class FaultGuardsCPU(unittest.TestCase):
                             ("slot_reservoir", (12, 4, 20, 21))):
            with self.subTest(mode=mode, bounds=bounds):
                with self.assertRaises(ValueError):
-                    implementation.FixedPatchStepper(cfg, context=context, coarse_boundary=mode,
-                                                     refinements=[bounds], level_blocks=[(1, 1)] * 2)
+                    implementation.FixedPatchStepper(cfg, bounds, context=context, coarse_boundary=mode)
                context.activate.assert_not_called()
        # Both exact channel limits and the original periodic two-cell margin
        # must reach context activation, with no allocation or device required.
@@ -119,20 +102,16 @@ class FaultGuardsCPU(unittest.TestCase):
                             ("periodic", (2, 2, 6, 5))):
            with self.subTest(mode=mode, bounds=bounds):
                with self.assertRaisesRegex(RuntimeError, "validated before CUDA"):
-                    implementation.FixedPatchStepper(cfg, context=context, coarse_boundary=mode,
-                                                     refinements=[bounds], level_blocks=[(1, 1)] * 2)
+                    implementation.FixedPatchStepper(cfg, bounds, context=context, coarse_boundary=mode)

    def fake_backend(self):
-        backend = FixedPatchStepper.__new__(FixedPatchStepper)
-        layout = BlockLayout2D.build(1, 1, 1, 1, periodic=(False, False))
-        backend._layouts = [layout, layout]  # blocked storage on both levels
-        backend._ddf_shifting = False  # unshifted on every level, as _check_storage expects
-        backend._level_cfgs = [LBMConfig(nx=16, ny=16, viscosity=.1)] * 2  # ddf_shifting off
+        backend = PeriodicPatchStepper.__new__(PeriodicPatchStepper)
+        backend.fine_layout = None  # dense fine storage
        backend.ctx = Mock(activate=nullcontext)
        backend._lock = RLock()
        backend._status = "ready"
        backend._stream = Mock()
-        backend._levels = [Mock(shape=(9, 1, 1), layout=layout), Mock(shape=(9, 1, 1), layout=layout)]
+        backend._levels = [Mock(shape=(9, 1, 1), layout=None), Mock(shape=(9, 1, 1), layout=None)]
        backend._allocations = [Mock(), Mock()]
        return backend

@@ -175,35 +154,20 @@ class FaultGuardsCPU(unittest.TestCase):
        context = Mock(activate=nullcontext, sm_arch="sm_70")
        with patch.object(implementation, "HALO", 2):
            with self.assertRaisesRegex(RuntimeError, "four-cell"):
-                FixedPatchStepper(cfg, context=context, refinements=[(4, 4, 12, 12)],
-                                     level_blocks=[(16, 16), (24, 24)])
+                PeriodicPatchStepper(cfg, (4, 4, 12, 12), context=context)
        p = implementation.FixedPatch2D.from_config(cfg, (4, 4, 12, 12))
-        from CelerisLab.lbm import hierarchy
-        bad = patch.object(implementation.FixedPatch2D, "from_config",
-                           return_value=replace(p, tau_coarse=.9))
-        # A pair tau inconsistent with its level config is rejected twice over, both before
-        # any device allocation or upload: first by the hierarchy chain check, and, if that
-        # is bypassed, by the per-level omega/tau check in _build_level.
-        with patch.object(cuda, "Stream"), patch.object(cuda, "mem_get_info", return_value=(1 << 40, 1 << 40)), bad, patch.object(cuda, "mem_alloc") as alloc, \
-                patch.object(cuda, "memcpy_htod") as upload:
-            with self.assertRaisesRegex(RuntimeError, "config chain is inconsistent"):
-                FixedPatchStepper(cfg, context=context, refinements=[p.bounds],
-                                     level_blocks=[(16, 16), (24, 24)])
-            with patch.object(hierarchy, "_check_chain"), \
-                    self.assertRaisesRegex(RuntimeError, "omega/tau"):
-                FixedPatchStepper(cfg, context=context, refinements=[p.bounds],
-                                     level_blocks=[(16, 16), (24, 24)])
-            alloc.assert_not_called()
-            upload.assert_not_called()
+        with patch.object(cuda, "Stream"), patch.object(implementation.FixedPatch2D, "from_config",
+                                                       return_value=replace(p, tau_coarse=.9)):
+            with self.assertRaisesRegex(RuntimeError, "omega/tau"):
+                PeriodicPatchStepper(cfg, p.bounds, context=context)
        module = Mock()
        module.get_global.return_value = (Mock(), 16)
-        with patch.object(cuda, "Stream"), patch.object(cuda, "mem_get_info", return_value=(1 << 40, 1 << 40)), patch.object(compiler, "generate_config"), \
+        with patch.object(cuda, "Stream"), patch.object(compiler, "generate_config"), \
                patch.object(compiler, "compile_kernel"), patch.object(compiler, "cleanup_build_dir"), \
                patch.object(compiler, "load_module", return_value=module), \
                patch.object(cuda, "memcpy_htod") as upload:
            with self.assertRaisesRegex(RuntimeError, "layout mismatch"):
-                FixedPatchStepper(cfg, context=context, refinements=[p.bounds],
-                                     level_blocks=[(16, 16), (24, 24)])
+                PeriodicPatchStepper(cfg, p.bounds, context=context)
            upload.assert_not_called()


@@ -271,27 +235,23 @@ class NormalBuildGPU(unittest.TestCase):
                parts = line.split()
                if len(parts) == 3 and parts[0] == "#define":
                    definitions[parts[1]] = parts[2]
-            generated.append((metadata["config"], definitions, metadata))
+            generated.append((metadata["config"], definitions))
            return build

        max_error = max_bc_error = 0.0
        with CudaContext(0) as context:
            with patch.object(compiler, "generate_config", side_effect=record_config):
-                backend = implementation.FixedPatchStepper(cfg, context=context, refinements=[bounds],
-                                                            coarse_boundary="slot_reservoir",
-                                                            level_blocks=whole_tiles(cfg, bounds))
+                backend = implementation.FixedPatchStepper(cfg, bounds, context=context,
+                                                            coarse_boundary="slot_reservoir")
            with backend:
                self.assertEqual(len(generated), 2)
-                fine_config, fine_defines, _ = generated[1]
-                # Blocked kernels see one tile window including the 1-cell halo.
-                self.assertEqual((fine_config["nx"], fine_config["ny"]), (fx+2, fy+2))
+                fine_config, fine_defines = generated[1]
+                self.assertEqual((fine_config["nx"], fine_config["ny"]), (fx, fy))
                for name in ("Y_WALL_BC", "TOP_WALL_BC", "WALL_SLOT_WIDTH",
                             "WALL_NORMAL_VELOCITY", "TOP_WALL_NORMAL_VELOCITY"):
                    self.assertEqual(float(fine_defines[name].removesuffix("f")), 0, name)
-                # The root's physical slot/reservoir walls come from BLOCK_ROOT_BC=1
-                # (module constant d_block_wall); the fine level has none.
-                root_meta, fine_meta = generated[0][2], generated[1][2]  # read before cleanup
-                self.assertEqual((root_meta["block_root_bc"], fine_meta["block_root_bc"]), (1, 0))
+                self.assertEqual(int(generated[0][1]["Y_WALL_BC"]), 2)
+                self.assertEqual(int(generated[0][1]["TOP_WALL_BC"]), 1)
                # Independently derive all target links, including the eight
                # whose coarse sources are uncovered, from physical midpoints.
                targets, corners = set(), set()
@@ -311,7 +271,9 @@ class NormalBuildGPU(unittest.TestCase):
                    expected = scalar_interval(*expected, bounds, slot=(amplitude, width))
                    backend.step()
                    self.assertEqual(backend.step_count, step)
-                    raw = owner_arrays(backend)
+                    raw = [np.empty(a.shape, np.float32) for a in initial]
+                    for level, array in zip(backend._levels, raw):
+                        cuda.memcpy_dtoh(array, level.current)
                    # Read raw storage before masking: active NaNs cannot be
                    # hidden by the public download's non-leaf cleanup.
                    self.assertTrue(np.isfinite(raw[0][:, coarse_leaf]).all())
@@ -348,9 +310,8 @@ class NormalBuildGPU(unittest.TestCase):
        state = ref.State(*(a * (1 + 1e-3*rng.standard_normal(a.shape))
                            for a in (initial.coarse_f, initial.fine_f)))
        max_pop = max_drift = max_moment = 0.0
-        with CudaContext(0) as context, FixedPatchStepper(
-                LBMConfig(nx=16, ny=16, viscosity=.1), context=context,
-                refinements=[(4, 4, 12, 12)], level_blocks=[(16, 16), (24, 24)]) as backend:
+        with CudaContext(0) as context, PeriodicPatchStepper(
+                LBMConfig(nx=16, ny=16, viscosity=.1), (4, 4, 12, 12), context=context) as backend:
            backend.initialize(*postcollision(state, spec, maps))
            before = totals(backend.download(), maps)
            for step in range(1, 9):
@@ -381,17 +342,17 @@ class NormalBuildGPU(unittest.TestCase):
        fine_mask[4:-4, 4:-4] = True
        initial[1][:, ~fine_mask] = np.nan
        expected = scalar_interval(*[a.astype(np.float64) for a in initial], bounds)
-        with CudaContext(0) as context, FixedPatchStepper(cfg, context=context, refinements=[bounds],
-                                                             level_blocks=[(13, 17), (14, 16)]) as backend:
+        with CudaContext(0) as context, PeriodicPatchStepper(cfg, bounds, context=context) as backend:
            self.assertEqual(backend.patch.fine_shape, (14, 16))
            p = backend.patch
-            allocated = 2 * 4 * sum(int(np.prod(level.layout.blocked_shape)) for level in backend._levels) \
-                + sum(a.nbytes for a in (p.ghost_fine, p.ghost_coarse, p.coalesce_coarse, p.coalesce_fine))
+            allocated = 2 * 9 * (17*13+16*14) * 4 + sum(a.nbytes for a in
+                        (p.ghost_fine, p.ghost_coarse, p.coalesce_coarse, p.coalesce_fine))
            self.assertLess(allocated, 128*1024)
            backend.initialize(*initial)
            backend.step()
            coarse, fine = backend.download()
-            raw_fine = owner_arrays(backend)[1]
+            raw_fine = np.empty((9, 14, 16), np.float32)
+            cuda.memcpy_dtoh(raw_fine, backend._levels[1].current)
            max_error = 0.0
            for got, want in zip((coarse, raw_fine), expected):
                np.testing.assert_array_equal(np.isnan(got), np.isnan(want))
@@ -20,9 +20,9 @@ import pycuda.driver as cuda

 import CelerisLab
 from CelerisLab.config import LBMConfig
-from CelerisLab.cuda import compiler
+from CelerisLab.cuda import compiler_v2 as compiler
 from CelerisLab.cuda.context import CudaContext
-from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
+from CelerisLab.lbm.refinement_stepper import FixedPatchStepper, PeriodicPatchStepper

 ROOT = Path(__file__).resolve().parents[3]
 POP_ATOL = 3e-6
@@ -52,11 +52,12 @@ def totals(arrays, maps):
    return np.array([populations.sum(), populations @ ref.CX, populations @ ref.CY])


-def level_storage(level):
-    """Raw device storage of one blocked level as dense (9, ny, nx)."""
-    blocked = np.empty(level.layout.blocked_shape, np.float32)
-    cuda.memcpy_dtoh(blocked, level.current)
-    return level.layout.to_dense(blocked)
+def global_array(module, name, shape):
+    out = np.empty(shape, dtype=np.uint32)
+    ptr, size = module.get_global(name)
+    assert size == out.nbytes
+    cuda.memcpy_dtoh(out, ptr)
+    return out


 class CoupledGPU(unittest.TestCase):
@@ -81,15 +82,15 @@ class CoupledGPU(unittest.TestCase):
        expected = postcollision(next_state, spec, maps)
        arrivals = (next_state.coarse_f, next_state.fine_f)

-        cases = (("blocked_unshifted", False, dict(level_blocks=[(16, 16), (24, 24)])),
+        cases = (("dense_unshifted", False, {}),
                 ("blocked_shifted", True,
-                  dict(level_blocks=[(4, 4), (6, 6)])))
+                  dict(base_block=(4, 4), fine_block=(6, 6))))
        with CudaContext(0) as context:
            for label, shifted, storage in cases:
                with self.subTest(storage=label):
                    cfg = LBMConfig(nx=16, ny=16, viscosity=.0192,
                                    ddf_shifting=shifted)
-                    with FixedPatchStepper(cfg, context=context, refinements=[(4, 4, 12, 12)],
+                    with FixedPatchStepper(cfg, (4, 4, 12, 12), context=context,
                                           **storage) as backend:
                        self.assertAlmostEqual(backend.patch.tau_coarse, spec.tau_coarse)
                        self.assertAlmostEqual(backend.patch.tau_fine, spec.tau_fine)
@@ -124,7 +125,7 @@ class CoupledGPU(unittest.TestCase):
                              "wrong-phase controls PASS", flush=True)

    def test_cleanup_attempts_every_allocation(self):
-        backend = FixedPatchStepper.__new__(FixedPatchStepper)
+        backend = PeriodicPatchStepper.__new__(PeriodicPatchStepper)
        backend.ctx = Mock(activate=nullcontext)
        backend._lock = RLock()
        backend._status = "failed"
@@ -146,9 +147,8 @@ class CoupledGPU(unittest.TestCase):
        spec = ref.make_spec(collision="srt")
        maps = ref.build_maps(spec)
        with CudaContext(0) as context:
-            with FixedPatchStepper(LBMConfig(nx=16, ny=16, viscosity=.1),
-                                      context=context, refinements=[(4, 4, 12, 12)],
-                                      level_blocks=[(16, 16), (24, 24)]) as backend:
+            with PeriodicPatchStepper(LBMConfig(nx=16, ny=16, viscosity=.1),
+                                      (4, 4, 12, 12), context=context) as backend:
                backend.initialize(*postcollision(ref.make_state(spec), spec, maps))
                coarse, fine = backend._levels

@@ -164,75 +164,122 @@ class CoupledGPU(unittest.TestCase):
                for operation in (backend.download, backend.step):
                    with self.assertRaisesRegex(RuntimeError, "failed"):
                        operation()
-                raw = level_storage(fine)
+                raw = np.empty(fine.shape, np.float32)
+                cuda.memcpy_dtoh(raw, fine.current)
                self.assertTrue(np.isfinite(raw.ravel()[backend.patch.coalesce_fine]).all())
            self.assertFalse(context.is_closed)
        print("partial failure: two fine hops completed, no coarse step published; cleanup PASS", flush=True)

-    def eight_steps(self, backend, spec, maps):
-        """Rest and seeded nonequilibrium, 8 steps against the scalar reference."""
-        p = backend.patch
-        for case in ("rest", "nonequilibrium_seed1"):
-            with self.subTest(case=case):
-                state = ref.make_state(spec, velocity=(0, 0) if case == "rest" else (.03, -.02))
-                if case != "rest":
-                    rng = np.random.default_rng(1)
-                    state = ref.State(*(a * (1 + 1e-3 * rng.standard_normal(a.shape))
-                                        for a in (state.coarse_f, state.fine_f)))
-                initial = postcollision(state, spec, maps)
-                backend.initialize(*initial)
-                before = totals(backend.download(), maps)
-                max_error = max_drift = max_moment_error = max_corner = 0.0
-                for step in range(1, 9):
-                    state, diagnostic = ref.advance_coarse_once(spec, maps, state)
-                    backend.step()
-                    actual = backend.download()
-                    expected = postcollision(state, spec, maps)
-                    self.assertEqual(backend.step_count, step)
-                    p.check_synchronization(step, state.tick, phase="postcollision_g")
-                    for got, want in zip(actual, expected):
-                        np.testing.assert_array_equal(np.isnan(got), np.isnan(want))
-                        error = float(np.nanmax(np.abs(got - want)))
-                        self.assertLessEqual(error, POP_ATOL)
-                        max_error = max(max_error, error)
-                    drift = float(np.max(np.abs(totals(actual, maps) - before)))
-                    moment_error = float(np.max(np.abs(totals(actual, maps) - totals(expected, maps))))
-                    self.assertLessEqual(drift, MOMENT_ATOL)
-                    self.assertLessEqual(moment_error, MOMENT_ATOL)
-                    max_drift, max_moment_error = max(max_drift, drift), max(max_moment_error, moment_error)
-                    f = backend._levels[1]
-                    raw = level_storage(f)
-                    np.testing.assert_array_equal(np.isfinite(raw), diagnostic["hop2_valid"])
-                    self.assertTrue(np.isfinite(raw.ravel()[p.coalesce_fine]).all())
-                    self.assertEqual(len(diagnostic["corner_packets"]), 8)
-                    for packet in diagnostic["corner_packets"].values():
-                        sx, sy = packet["source_coarse_xy"]
-                        self.assertTrue(maps.coarse_active[sy, sx])
-                        values = [raw[packet["direction"], y, x]
-                                  for y, x in packet["children_fine_yx"]]
-                        error = float(np.max(np.abs(np.array(values) - packet["packets"])))
-                        self.assertLessEqual(error, POP_ATOL)
-                        max_corner = max(max_corner, error)
-                print(f"{case}: steps=8 max_population={max_error:.9g} "
-                      f"max_leaf_drift={max_drift:.9g} max_leaf_error={max_moment_error:.9g} "
-                      f"max_corner_packet={max_corner:.9g}; PASS", flush=True)
-
-    def test_bounded_equivalence_blocked(self):
-        """The same eight-step reference and launch-failure checks on whole-level tiles."""
+    def test_bounded_equivalence_and_ownership(self):
+        self.assertEqual(Path(CelerisLab.__file__).resolve(), ROOT / "src/CelerisLab/__init__.py")
+        self.assertEqual(Path(compiler._KERNEL_DIR).resolve(), ROOT / "src/CelerisLab/lbm/kernels")
        spec = ref.make_spec(collision="srt")
        maps = ref.build_maps(spec)
        cfg = LBMConfig(nx=16, ny=16, viscosity=.1)
+        compile_original = compiler.compile_kernel
+
+        def audit_compile(*args, **kwargs):
+            # Only this private build's entry differs. Production source unchanged.
+            entry = Path(kwargs["build_dir"]) / "refinement_audit.cu"
+            entry.write_text('#define CELERIS_REFINEMENT_AUDIT\n#include "refinement.cu"\n')
+            kwargs["source"] = entry.name
+            return compile_original(*args, **kwargs)
+
        with CudaContext(0) as context:
-            with FixedPatchStepper(cfg, context=context, refinements=[(4, 4, 12, 12)],
-                                      level_blocks=[(16, 16), (24, 24)]) as backend:
-                self.eight_steps(backend, spec, maps)
+            with patch.object(compiler, "compile_kernel", audit_compile):
+                backend = PeriodicPatchStepper(cfg, (4, 4, 12, 12), context=context)
+            with backend:
+                # Production allocations: four DDF arrays and four compact tables.
+                device_bytes = 2 * 9 * (16**2 + 24**2) * 4
+                p = backend.patch
+                device_bytes += sum(a.nbytes for a in (p.ghost_fine, p.ghost_coarse,
+                                                       p.coalesce_coarse, p.coalesce_fine))
+                audit_bytes = sum((2 + 9) * np.prod(level.shape[1:]) * 4
+                                  for level in backend._levels)
+                self.assertLess(device_bytes + audit_bytes, 128 * 1024)
+                print(f"device_arrays={device_bytes} audit_arrays={audit_bytes} bytes", flush=True)
+                for case in ("rest", "nonequilibrium_seed1"):
+                    with self.subTest(case=case):
+                        state = ref.make_state(spec, velocity=(0, 0) if case == "rest" else (.03, -.02))
+                        if case != "rest":
+                            rng = np.random.default_rng(1)
+                            state = ref.State(*(a * (1 + 1e-3 * rng.standard_normal(a.shape))
+                                                for a in (state.coarse_f, state.fine_f)))
+                        initial = postcollision(state, spec, maps)
+                        backend.initialize(*initial)
+                        for level in backend._levels:
+                            for name in ("collision_counts", "explosion_counts", "incoming_counts"):
+                                ptr, size = level.module.get_global(name)
+                                cuda.memset_d8(ptr, 0, size)
+                        before = totals(backend.download(), maps)
+                        max_error = max_drift = max_moment_error = max_corner = 0.0
+                        for step in range(1, 9):
+                            state, diagnostic = ref.advance_coarse_once(spec, maps, state)
+                            backend.step()
+                            actual = backend.download()
+                            expected = postcollision(state, spec, maps)
+                            self.assertEqual(backend.step_count, step)
+                            p.check_synchronization(step, state.tick, phase="postcollision_g")
+                            for got, want in zip(actual, expected):
+                                np.testing.assert_array_equal(np.isnan(got), np.isnan(want))
+                                error = float(np.nanmax(np.abs(got - want)))
+                                self.assertLessEqual(error, POP_ATOL)
+                                max_error = max(max_error, error)
+                            drift = float(np.max(np.abs(totals(actual, maps) - before)))
+                            moment_error = float(np.max(np.abs(totals(actual, maps) - totals(expected, maps))))
+                            self.assertLessEqual(drift, MOMENT_ATOL)
+                            self.assertLessEqual(moment_error, MOMENT_ATOL)
+                            max_drift, max_moment_error = max(max_drift, drift), max(max_moment_error, moment_error)
+                            c, f = backend._levels
+                            np.testing.assert_array_equal(global_array(c.module, "collision_counts", (16, 16)),
+                                                          step * maps.coarse_active)
+                            np.testing.assert_array_equal(global_array(f.module, "collision_counts", (24, 24)),
+                                                          2 * step * maps.fine_active)
+                            np.testing.assert_array_equal(global_array(f.module, "explosion_counts", (24, 24)),
+                                                          step * ~maps.fine_active)
+                            owners = np.zeros((9, 16, 16), np.uint32)
+                            owners.ravel()[p.coalesce_coarse] = step
+                            np.testing.assert_array_equal(global_array(c.module, "incoming_counts", owners.shape), owners)
+                            raw = np.empty(f.shape, np.float32)
+                            cuda.memcpy_dtoh(raw, f.current)
+                            np.testing.assert_array_equal(np.isfinite(raw), diagnostic["hop2_valid"])
+                            self.assertTrue(np.isfinite(raw.ravel()[p.coalesce_fine]).all())
+                            self.assertEqual(len(diagnostic["corner_packets"]), 8)
+                            for packet in diagnostic["corner_packets"].values():
+                                sx, sy = packet["source_coarse_xy"]
+                                self.assertTrue(maps.coarse_active[sy, sx])
+                                values = [raw[packet["direction"], y, x]
+                                          for y, x in packet["children_fine_yx"]]
+                                error = float(np.max(np.abs(np.array(values) - packet["packets"])))
+                                self.assertLessEqual(error, POP_ATOL)
+                                max_corner = max(max_corner, error)
+                        print(f"{case}: steps=8 max_population={max_error:.9g} "
+                              f"max_leaf_drift={max_drift:.9g} max_leaf_error={max_moment_error:.9g} "
+                              f"max_corner_packet={max_corner:.9g}; ownership/counts PASS", flush=True)
+
+                # Fail before any additional time step; no stale synchronized export.
                with patch.object(backend._pairs[0], "explode", side_effect=RuntimeError("injected launch failure")):
                    with self.assertRaisesRegex(RuntimeError, "injected"):
                        backend.step()
-                for operation in (backend.download, backend.step):
-                    with self.assertRaisesRegex(RuntimeError, "failed"):
-                        operation()
+                with self.assertRaisesRegex(RuntimeError, "failed"):
+                    backend.download()
+                with self.assertRaisesRegex(RuntimeError, "failed"):
+                    backend.step()
            self.assertFalse(context.is_closed)
+            with self.assertRaisesRegex(RuntimeError, "closed"):
+                backend.download()
+
+            # Compile and execute the actual non-audited entry as well: one step.
+            with PeriodicPatchStepper(cfg, (4, 4, 12, 12), context=context) as plain:
+                rest = postcollision(ref.make_state(spec), spec, maps)
+                plain.initialize(*rest)
+                plain.step()
+                for got, expected in zip(plain.download(), rest):
+                    np.testing.assert_allclose(got, expected, rtol=0, atol=POP_ATOL, equal_nan=True)
+                print("normal-build rest: steps=1 PASS; source/kernel paths pinned", flush=True)
+        rss = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss / 1024
+        print(f"host_peak_rss={rss:.3f} MiB", flush=True)
+

 if __name__ == "__main__":
    unittest.main(verbosity=2)
@@ -3,13 +3,11 @@
 Jets use the same coarse cell-edge frame as fine_objects. The equivalent jet
 of Astra's actuated fixture is SurfaceActuation(0.7, 0.85, U).as_jet((16, 12),
 2.15). Frozen GPU gates (8 coarse steps, both coarse boundary modes):
-  G1 jet run: 1x1/2x2/4x2 fine blocks (blocked (12, 8) root) bit-identical to the
-     FINE_BLOCKS[0] reference (values + NaN masks);
-  G2 jet vs legacy actuation (blocked reference): max |dG| over finite fine ACTIVE fluid
+  G1 jet run: dense vs 1x1/2x2/4x2 fine blocks bit-identical (values + NaN masks);
+  G2 jet vs legacy actuation (dense): max |dG| over finite fine ACTIVE fluid
     populations <= 1e-6 (velocity rows differ only by FP32 rounding; reported);
-  G3 amplitude change after 4 steps (+U -> -U), 4 more steps: blocked layouts
-     bit-identical to the FINE_BLOCKS[0] reference and the run differs from the
-     unchanged run.
+  G3 amplitude change after 4 steps (+U -> -U), 4 more steps: dense vs blocked
+     bit-identical and the run differs from the unchanged run.
 """
 from contextlib import nullcontext
 import unittest
@@ -18,7 +16,7 @@ from unittest.mock import Mock, patch
 import numpy as np
 import pycuda.driver as cuda

-from CelerisLab.cuda import compiler
+from CelerisLab.cuda import compiler_v2 as compiler
 from CelerisLab.lbm.descriptors import SOLID
 from CelerisLab.lbm.refinement import ACTIVE, FixedPatch2D, prepare_fine_boundary
 from CelerisLab.lbm.refinement_stepper import FixedPatchStepper
@@ -28,11 +26,6 @@ from test_refinement_blocked_fine import (BOUNDS, FINE_BLOCKS, NX, NY, STEPS, ac
 SPEED = .002


-def root_and(fine_block):
-    """Blocked levels with a (12, 8) root tile."""
-    return [(12, 8), fine_block]
-
-
 def jet(speed=SPEED):
    return actuation(speed).as_jet((16.0, 12.0), 2.15)

@@ -43,14 +36,13 @@ def mocked(objects, jets, fine_block):
    module = Mock(get_function=Mock(return_value=Mock()))
    module.get_global.return_value = (Mock(), 20)
    htod = Mock()
-    with patch.object(cuda, "Stream"), patch.object(cuda, "mem_get_info", return_value=(1 << 40, 1 << 40)), patch.object(compiler, "generate_config"), \
+    with patch.object(cuda, "Stream"), patch.object(compiler, "generate_config"), \
            patch.object(compiler, "compile_kernel"), patch.object(compiler, "cleanup_build_dir"), \
            patch.object(compiler, "load_module", return_value=module), \
            patch.object(cuda, "mem_alloc", side_effect=lambda size: Mock()), \
            patch.object(cuda, "memcpy_htod", htod):
-        backend = FixedPatchStepper(config("periodic"), context=context, refinements=[BOUNDS],
-                                    level_blocks=root_and(fine_block), level_objects=[(), tuple(objects)],
-                                    level_jets=[(), tuple(jets)])
+        backend = FixedPatchStepper(config("periodic"), BOUNDS, context=context,
+                                    fine_objects=objects, fine_block=fine_block, jets=jets)
    return backend, htod


@@ -107,7 +99,7 @@ class JetStepperCPU(unittest.TestCase):
    def test_blocked_amplitude_update_rewrites_only_velocity_column(self):
        backend, htod = mocked([circle()], [jet()], FINE_BLOCKS[1])
        before = backend._fine_tables["surface_velocity"].copy()
-        device = backend._device_boundaries[1].device["surface_velocity"]
+        device = backend._link_tables["surface_velocity"]
        with patch.object(cuda, "memcpy_htod", htod):
            htod.reset_mock()
            backend.set_jet_amplitude(0, -SPEED)
@@ -123,7 +115,7 @@ class JetStepperCPU(unittest.TestCase):
        backend.close()

    def test_copy_failure_marks_failed_and_keeps_host_amplitude(self):
-        for fine_block in (FINE_BLOCKS[1],):  # blocked tables
+        for fine_block in (None, FINE_BLOCKS[1]):  # dense SoA and blocked tables
            backend, _ = mocked([circle()], [jet()], fine_block)
            status = backend._status
            for bad, error in ((1.5, ValueError), (float("nan"), ValueError)):
@@ -133,7 +125,9 @@ class JetStepperCPU(unittest.TestCase):
            with self.assertRaises(IndexError):
                backend.set_jet_amplitude(1, -SPEED)
            self.assertEqual(backend._status, status)
-            before = backend._fine_tables["surface_velocity"].copy()
+            dense = backend._fine_curved
+            before = (dense.surface_velocity if dense is not None
+                      else backend._fine_tables["surface_velocity"]).copy()
            boom = Mock(side_effect=RuntimeError("copy failed"))
            with patch.object(cuda, "memcpy_htod", boom), patch.object(cuda, "memcpy_htod_async", boom), \
                    patch.object(cuda, "pagelocked_empty", side_effect=lambda shape, dtype: np.empty(shape, dtype)):
@@ -141,7 +135,11 @@ class JetStepperCPU(unittest.TestCase):
                    backend.set_jet_amplitude(0, -SPEED)
            self.assertEqual(backend._status, "failed", fine_block)
            self.assertEqual(backend._jets[0].amplitude, SPEED)
-            np.testing.assert_array_equal(backend._fine_tables["surface_velocity"], before)
+            if dense is not None:
+                self.assertEqual(dense.jets[0].amplitude, SPEED)
+                np.testing.assert_array_equal(dense.surface_velocity, before)
+            else:
+                np.testing.assert_array_equal(backend._fine_tables["surface_velocity"], before)
            with self.assertRaisesRegex(RuntimeError, "failed"):
                backend.step(1)
            with self.assertRaisesRegex(RuntimeError, "failed"):
@@ -151,10 +149,8 @@ class JetStepperCPU(unittest.TestCase):
    def test_jet_overlapping_legacy_aperture_rejected_before_cuda(self):
        with patch.object(cuda, "Stream", side_effect=AssertionError("CUDA touched")):
            with self.assertRaisesRegex(ValueError, "SurfaceActuation"):
-                FixedPatchStepper(config("periodic"), context=Mock(), refinements=[BOUNDS],
-                                  level_blocks=root_and(FINE_BLOCKS[1]),
-                                  level_objects=[(), (circle(actuation(SPEED)),)],
-                                  level_jets=[(), (jet(),)])
+                FixedPatchStepper(config("periodic"), BOUNDS, context=Mock(),
+                                  fine_objects=[circle(actuation(SPEED))], jets=[jet()])


 def run(mode, fine_block, *, objects, jets, switch=None):
@@ -162,9 +158,8 @@ def run(mode, fine_block, *, objects, jets, switch=None):
    cfg = config(mode)
    patch_ = FixedPatch2D.from_config(cfg, BOUNDS)
    with CudaContext(0) as context, FixedPatchStepper(
-            cfg, context=context, coarse_boundary=mode, refinements=[BOUNDS],
-            level_blocks=root_and(fine_block), level_objects=[(), tuple(objects)],
-            level_jets=[(), tuple(jets)]) as backend:
+            cfg, BOUNDS, context=context, coarse_boundary=mode, fine_objects=objects,
+            fine_block=fine_block, jets=jets) as backend:
        backend.initialize(state(NY, NX, 1), state(*patch_.fine_shape, 2))
        if switch is None:
            backend.step(STEPS)
@@ -181,29 +176,29 @@ class JetStepperGPU(unittest.TestCase):
            patch_, boundary = fine_circle(mode)
            fluid = (patch_.fine_kind == ACTIVE) & (
                (boundary.flags.reshape(patch_.fine_shape) & SOLID) == 0)
-            reference = run(mode, FINE_BLOCKS[0], objects=[circle()], jets=[jet()])
-            self.assertTrue(np.isfinite(reference[1][1][:, fluid]).all(), mode)
-            for fine_block in FINE_BLOCKS[1:]:
+            dense = run(mode, None, objects=[circle()], jets=[jet()])
+            self.assertTrue(np.isfinite(dense[1][1][:, fluid]).all(), mode)
+            for fine_block in FINE_BLOCKS:
                public, raw_arrays = run(mode, fine_block, objects=[circle()], jets=[jet()])
-                for a, b in zip(public + tuple(raw_arrays), reference[0] + tuple(reference[1])):
+                for a, b in zip(public + tuple(raw_arrays), dense[0] + tuple(dense[1])):
                    np.testing.assert_array_equal(a, b, err_msg=f"G1 {mode} {fine_block}")
-            legacy = run(mode, FINE_BLOCKS[0], objects=[circle(actuation(SPEED))], jets=())
-            diff = np.abs(reference[1][1][:, fluid] - legacy[1][1][:, fluid]).max()
+            legacy = run(mode, None, objects=[circle(actuation(SPEED))], jets=())
+            diff = np.abs(dense[1][1][:, fluid] - legacy[1][1][:, fluid]).max()
            print(f"G2 {mode}: jet vs legacy fine max |dG| = {diff:.3e}", flush=True)
            self.assertLessEqual(diff, 1e-6, mode)

    def test_g3_amplitude_switch_layouts_identical(self):
        for mode in ("periodic", "slot_reservoir"):
-            reference = run(mode, FINE_BLOCKS[0], objects=[circle()], jets=[jet()], switch=-SPEED)
-            steady = run(mode, FINE_BLOCKS[0], objects=[circle()], jets=[jet()])
-            self.assertFalse(np.array_equal(reference[1][1], steady[1][1], equal_nan=True), mode)
-            for fine_block in FINE_BLOCKS[1:]:
+            dense = run(mode, None, objects=[circle()], jets=[jet()], switch=-SPEED)
+            steady = run(mode, None, objects=[circle()], jets=[jet()])
+            self.assertFalse(np.array_equal(dense[1][1], steady[1][1], equal_nan=True), mode)
+            for fine_block in FINE_BLOCKS:
                public, raw_arrays = run(mode, fine_block, objects=[circle()], jets=[jet()],
                                         switch=-SPEED)
-                for a, b in zip(public + tuple(raw_arrays), reference[0] + tuple(reference[1])):
+                for a, b in zip(public + tuple(raw_arrays), dense[0] + tuple(dense[1])):
                    np.testing.assert_array_equal(a, b, err_msg=f"G3 {mode} {fine_block}")
-            print(f"G3 {mode}: amplitude switch, {len(FINE_BLOCKS) - 1} layouts "
-                  f"identical to {FINE_BLOCKS[0]}", flush=True)
+            print(f"G3 {mode}: amplitude switch dense vs {len(FINE_BLOCKS)} layouts identical",
+                  flush=True)


 if __name__ == "__main__":
--- a/Show More
+++ b/Show More