refinement.cu keeps its header, includes, the curved_double.cu entry
block, the #error guard and the optional CELERIS_REFINEMENT_AUDIT
counters, then includes step/patch_step.cu, which holds patch_collide,
PatchExplode, PatchFineStep and PatchCoarseStep moved verbatim (checked
against the previous refinement.cu) in the same order.
test_refinement_mrt reads refinement.cu + step/patch_step.cu; its
assertions are unchanged. The audit-build tests include refinement.cu
from the build root, so step/patch_step.cu still resolves.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
blocks.cu keeps its header, includes and #error guards and then includes
the kernels, as kernel_v2.cu does:
- step/block_transfer.cu: BlockHaloCopy, BlockExplode,
BlockExplodeProjected, BlockCoalesce
- step/block_boundary.cu: BlockCurvedBoundary, BlockAccumulateExchange
- step/block_step.cu: BlockStep (rejects BLOCK_ROOT_BC == 2)
- step/block_channel.cu: unchanged (mode 2 only)
Kernel text moved verbatim (checked against the previous blocks.cu) and in
the same definition order, so the translation unit is the same sequence of
kernels. The two source-reading tests follow the move:
test_projected_explosion reads step/block_transfer.cu, test_refinement_mrt
reads blocks.cu + step/block_step.cu; their assertions are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Horsea's gates showed block_ddf/block_slot change real PTX in
BlockHaloCopy, BlockExplode and BlockExplodeProjected (6/10 modules;
BlockExplodeProjected 38 -> 40 registers) although the results stayed
exact. The stack's bar is byte-identical PTX for naming-only steps, so
those three kernels return to their K3 expressions, the helpers leave
core/layout.cuh, and test_projected_explosion goes back to its K3 form.
fill_invalid stays (its kernels were byte-identical). K4 vs K3 is now
only the NaN-fill helper.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
K4 made BlockExplodeProjected address storage through block_ddf and
block_slot (core/layout.cuh). This host test compiles the kernel text
extracted from blocks.cu, so it now also extracts those two helper
definitions from core/layout.cuh (device qualifiers -> inline) instead of
failing to compile. Assertions and tolerances unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
operators/helpers.cuh already had bounce_back_swap (the uniform steps use
it). It moves verbatim, #pragma unroll included, into its own
operators/bounce_back_swap.cuh, which helpers.cuh includes; the channel
root includes that header and calls it. bounce_back_swap_d2q9 from the
previous commit is removed. bounce_back.cuh keeps only the half-way rows
helper.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- boundary/bounce_back.cuh: apply_wall_bb_d2q9_rows(y, top_row, ...) is
the existing half-way y-wall correction with an explicit top row;
apply_wall_bb_d2q9_y_pull keeps its signature and calls it with NY - 2.
bounce_back_swap_d2q9(f) names the full opposite-pair swap.
- streaming/pull_double_buffer.cuh: load_node(fi, n, f) loads one node's
stored populations.
- step/block_channel.cu uses all three instead of its inline copies: the
wall swap, the outlet donor gather, and rows 1 / ny-2 (now one call with
the physical top row; same y == 1 / y == top_row order and loads).
Same statements, branch order, indices and arithmetic.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- core/layout.cuh: block_ddf(f, block_cell[, cells]) returns the start of a
block's populations and block_slot(block_cell, cells, q) the offset of one
population; BlockHaloCopy, BlockExplode and BlockExplodeProjected use them
instead of repeating (c / cells) * NQ * cells [+ q * cells + c % cells].
BlockExplode keeps its loop-hoisted parent base; types stay uint64_t /
unsigned long; the expressions are the same ones, now named.
- core/cell_kind.cuh: fill_invalid(f) writes the NaN marker; BlockStep,
BlockChannelStep and PatchFineStep call it instead of three copies of the
loop.
No change to operation order, types, launches or arithmetic.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- core/cell_kind.cuh holds KIND_ACTIVE/INNER/OUTER (was a constexpr in
blocks.cu that step/block_channel.cu and the channel_root.cpp shim
relied on or redefined); blocks.cu, block_channel.cu and the shim now
include it.
- blocks.cu checks the lattice/model #error guard before the
BLOCK_ROOT_BC guard and mode includes, so an unsupported build reports
the real reason first.
- operators/block_collision.cuh includes what it uses (params, macro,
SRT/MRT collision) instead of relying on blocks.cu's include order.
Preprocessor-only: same kernels, same statements, same arithmetic.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Manifests and receipts unchanged. Each of the 8 group-A identity tests now
extracts its manifest's recorded commit and verifies hashes there instead of
against the current checkout. Tamper assertions and runner source refusals are
kept; missing recorded history is an explicit failure, not a skip or repin.
Names the exact SRT/MRT parity tests and their sub-case scope (SRT: one
of 18 sub-cases, unidentified; MRT: 12 sub-cases, default outlet settings,
with the runs that recorded the pass), cites blocks.py:134 for tile
divisibility, labels refinement.py:457 as the fine-boundary re-check, and
says 'measured at 36ffc8b1' for the Taylor-Green observation.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Test-side only; no production, manifest or pin change. Group A identity
checks (test_source_pin and the jets.py pins) are unchanged and stay
failing as recorded.
B test_refined_dfg: receipt/plan fixtures load the frozen manifest JSON
that verify_source returns for a matching tree (frozen_manifest), and
the two manifest-selection assertions call verify_source with only its
working-tree hash comparison neutralised (selected_manifest), so the
malformed/collision/observation-mode/checkout receipt checks run again.
test_source_pin remains the single identity check.
C test_channel_root: int(owner)-1 (uint64-int became float on NumPy 1);
test_shifted_storage: FP32 NaN fill so the arrays stay FP32. Both reach
the intended immutable/representation/FP32/density checks.
D test_block_stepper fake gets _ddf_shifting; test_fine_surface_gpu sets
_boundary_hosts/_level_jets instead of the read-only fine_boundary view;
test_refinement_followup fake_backend sets _layouts/_ddf_shifting/
_level_cfgs instead of the read-only fine_layout view (as reviewed in
32da2d73).
E test_kernel_contracts_reject_before_upload: an inconsistent pair tau is
now first rejected by the hierarchy chain check; the test asserts that
rejection, then bypasses _check_chain to prove the per-level omega/tau
guard still rejects, and asserts no mem_alloc/memcpy_htod in either case.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Copy tests/integration/test_checkpoint_bodies_gpu.py byte-for-byte from 34b30c434f. Base c95a7baae4. No production or dependency changes. GPU execution is deferred to the assigned executor.
One page listing supported settings, each with the guard enforcing it
(file:line at c95a7ba) and the existing tests exercising it. Measured
observations are labelled by their exact tested case (SRT strict-parity
1/4320 values ~2 ULP with unresolved cause; V3 TG constant explosion
aggregate; curved-wall r1 pilot; V4 NOT_STEADY), not stated as general
bounds. Documentation only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The pinned BUILD_SOURCES are the a135252 qualification build inputs.
generate_config hashes the whole kernel snapshot, so after K1/K2 the
working-tree kernels (legitimately) no longer reproduce them. The test
now archives src/CelerisLab/lbm/kernels at NUMERICAL_REVISION and points
compiler_v2._KERNEL_DIR there; config headers are still rendered by
today's compiler. BUILD_SOURCES and the equality assertion are unchanged;
each size is a subTest; a missing NUMERICAL_REVISION fails (no skip).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
surface_link_velocity already runs jet_link_shapes; CurvedLinkSoA.
assign_surface and _prepare_blocked_boundary then ran it again with the
same jets and hits to fill the host jet cache. A private _surface_links
returns (velocity, aperture, shape, jet_id) from the one call; both
callers use it. surface_link_velocity keeps its two-value public
contract (now a thin wrapper). Same values, same validation order, same
cache contents; host geometry preparation only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Drop legacy constructor prose (child=/mutual exclusion), retarget
README and fixed_three_level_multipatch examples to refinements /
level_blocks, and fix stepper error/doc strings that still named
fine_block/base_block or a fixed gap >= 4 (unions use dependency
closure). Wording only; no behavior change.
Removes the legacy entry points now that every caller is migrated:
FixedPatchStepper(cfg, *, context, refinements, coarse_boundary,
level_blocks, level_objects, level_jets, observe_*, explosion) and
initialize(*level_g). Deleted: positional bounds, fine_block, base_block,
fine_objects, jets, child, child_block, child_objects, child_jets,
child_rectangles_in_parent, initialize(coarse_g/fine_g/child_g),
_legacy_levels and _initial_levels. Docstring, README and docs describe
the level lists. Compatibility views (patch, fine_boundary, ...) and the
PeriodicPatchStepper alias are unchanged (separate follow-up).
Error -> rule table (legacy keyword check -> level-list rule now firing):
- several patches without fine_block/base_block -> _check_levels
'several patches require fine_block and base_block' (unchanged rule)
- base_block without fine_block -> 'a blocked level requires blocked
refined levels below it'
- nested child without all blocks -> 'three or more levels must all be
blocked' / 'a blocked level requires ...'
- fine_objects/jets with a nested child ('only on L2') -> 'bodies/Jets are
supported only on the finest level of a nested hierarchy'
- child_objects/child_jets without a nested child -> 'level_objects/
level_jets needs one entry per level'
- scalar child with several L1 patches ('scalar child requires one parent
patch') -> hierarchy 'a scalar child requires a scalar parent'
- child vs child_rectangles_in_parent mutually exclusive -> no level-list
equivalent (cannot be expressed); covered by the new explicit
RemovedLegacyAPI test: removed keywords and positional bounds raise
TypeError before any context use, initialize has only *level_g.
The InitialLevels test of the removed adapter is replaced by that test.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
One local levels(bounds, blocks, objects, jets, **rest) builds the
two-level lists; run_steps and every rejection case use it. All five
before-CUDA rejections, the margin and one-jet checks keep their inputs.
Assertions unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Dense root + blocked fine level is level_blocks=[None, block]; bodies go
to level_objects[1]. Both before-CUDA rejections (non-dividing and
malformed fine block) keep their inputs. Assertions unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
FixedPatchStepper and the PeriodicPatchStepper alias are called with
refinements=[bounds] instead of the positional legacy bounds (same
_legacy_levels result). Assertions, including the before-CUDA and
four-cell/omega/layout rejections, unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
KW becomes levels() (scalar L1 + scalar L2, per-level blocks, objects,
Jets). All rejection cases are kept per level. Two cases change branch:
L2 bodies/Jets with no nested child (legacy 'child_objects/child_jets
require a nested child') are now an entry for a level that does not
exist ('needs one entry per level'); several L1 patches + scalar child
now meets 'a scalar child requires a scalar parent'. Both still reject
before CUDA. Assertions unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
test_refinement_mrt_gpu, test_shifted_jets_gpu and
test_refinement_blocked_base pass refinements/level_* lists. The three
blocked-base rejection cases keep their inputs per level; base-only
blocks now meet the level-list rule (blocked refined levels below) where
the legacy keyword check fired before. Assertions unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
test_exchange_window, test_raw_exchange_gpu (one fine(block) helper for
the repeated circle+jet construction), test_refinement_jets,
test_multiple_pinned_remap_gpu and test_pinned_body_remap_gpu pass
refinements/level_* lists. Dense fine levels stay level_blocks=[None,
None|block] as before. Assertions unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
test_refinement_gpu (incl. PeriodicPatchStepper alias calls),
test_channel_root_gpu, test_fine_surface_gpu, test_shifted_remap_gpu,
test_runtime_child_remap_gpu and test_refinement_three_level_multipatch_gpu
pass refinements/level_* lists; scalar and set children map to the second
refinement entry exactly as _legacy_levels did. Assertions unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
test_shifted_jets, test_shifted_storage, test_raw_exchange and
test_channel_root pass refinements/level_* lists; examples
fixed_patch_jet and fixed_periodic_patch pass refinements=[BOUNDS]
instead of the positional bounds. Assertions kept; the shifted-storage
case base_block-without-fine_block now expects the level-list rule
'blocked refined levels below' (same constraint).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
KW (legacy base/fine/child blocks + child_rectangles_in_parent) becomes
levels(): refinements=[PARENTS, CHILDREN], level_blocks, level_objects,
level_jets. Every negative case is kept and expressed per level (child
rectangles, per-level blocks); the L1-body case now expects the level-list
rule's message 'only on the finest level' instead of the legacy keyword
message 'only on L2' (same constraint).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The module-level import of CelerisLab.lbm.block_stepper loaded
pycuda.driver into the shared test process, unlike the other CPU host
tests (which load mocked copies via AST). find_spec gives the source path
without the import. Test content and assertions unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
run_refined_dfg, run_dfg_re20, time_window_dfg, time_refined_dfg,
run_jet_circle_cv and examples/fixed_patch_circle pass refinements /
level_blocks / level_objects / level_jets instead of bounds, base_block,
fine_block, fine_objects and jets. One-to-one with _legacy_levels: same
rectangle (FixedPatch2D frame), same blocks, bodies and Jets.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The only in-library caller of the legacy constructor keywords now passes
refinements=[L1 bounds, L2 set], level_blocks=[base, fine, child],
level_objects=[(), (), child_objects], level_jets=[(), (), jets] -- the
exact lists _legacy_levels produced for it. Legacy keywords stay until
every caller is migrated (final P1 commit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review follow-up (non-blocking items): a launch failure on the second of
three steps fails closed without the batch wait, keeps step_count at its
entry value and blocks download; the docstring now states, as for
FixedPatchStepper, that async faults surface at the single wait.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Same change as the reviewed FixedPatchStepper end-sync (cherry-picked
ad1d82c+512dcc9 above): the n steps are enqueued in order and the host
waits once; step(0) launches nothing; step_count advances after the wait
and keeps its entry value on failure. Launches, buffer roles and bits are
those of n step(1) calls. Mocked-driver host test covers the single
wait, launch/role equality with n step(1), step(0) and failure.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Enqueue all n base intervals in order on the one stream and synchronize
once before returning (step(0): no launch, no wait). step_count, the raw
exchange ledger and an open window are published only after that wait;
any failure leaves step_count at its entry value and invalidates both
ledgers. Async device errors are no longer attributed to an interval;
the docstring says to use step(1) loops to localize one. Launch order and
buffers are unchanged. Adds a mocked-driver host test of the contract.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
BlockStep (slot-reservoir wall rows) and BlockChannelStep computed the
global cell of a block-window cell with the same four lines.
block_global_xy (core/layout.cuh, 2D, __forceinline__) now holds them;
both kernels call it with their own domain width. Same integer
operations: no numerical, ABI or scheduling change. The slot-reservoir
kernel layout is unchanged (no split).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
BlockStep, BlockChannelStep and refinement.cu patch_collide each carried
the same compute_rho_u -> omega clamp -> collide_mrt_no_force sequence.
collide_mrt_unforced (collision_mrt.cuh, __forceinline__) now holds it
once; the three sites call it. Same operations in the same order: no
numerical, ABI or SRT change. The host source contract now requires one
helper definition and exactly one call per file, and forbids the old
inline sequence.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Grok