test: cover TRT restart and actual-rate refinement phase

Integrate reviewed test changes from 7c1517de819741907fb945ea0ab7864734fa3856 and 340ba4e1dc6386581839c0b28c8fe8a1f4aeb70b unchanged. Document bounded 2D implementation, evidence and open gates; production source and historical manifests are unchanged.

docs/validation/2d_scope.md

@@ -0,0 +1,51 @@
+# Bounded 2D implementation and validation scope
+
+This inventory uses dev `8403956a9afc0b0bb89265cec374392013d89016` as its
+production source checkpoint. The accompanying tests add coverage; their presence
+is not a passing GPU receipt. Historical campaigns retain their original source
+identities and acceptance gates.
+
+## Implementation boundaries
+
+- Uniform D2Q9 supports SRT, TRT and MRT. An available collision selector does not
+  qualify every wall, inlet/outlet, forcing, storage or LES combination.
+- Fixed refinement supports SRT/MRT, FP32 and double-buffer streaming; TRT and LES
+  are rejected. Shifted refinement requires all-blocked storage. Refined channel
+  boundaries have narrower configuration guards than uniform boundaries.
+- Uniform checkpoints support compatible same-method continuation. Changing
+  collision or LES parameters is a different evolution, not exact continuation.
+  Jet checkpoints and refined checkpoints are unsupported; synchronized remapping
+  is not persistent restart.
+- Keep these conclusions 2D. Local D3Q19 operator tests do not establish a supported
+  3D Simulation runtime.
+
+## Evidence and remaining gaps
+
+| Area | Available evidence | Boundary of the conclusion |
+|---|---|---|
+| MRT cylinder flow and observation windows | Full matched uniform/refined MRT DFG and the window-ledger campaign passed their frozen periodicity, reference and matched-comparison gates. The window campaign used `185a919`; dev adds only a docstring correction. | This does not qualify other collision/boundary combinations or every refinement interface. Timing improvements are not additional accuracy evidence. |
+| Uniform forced channel | Six Poiseuille GPU cases on `8403956`: SRT/TRT/MRT at wall-normal resolutions H16/H32 passed. | This is the tested analytic channel configuration, not open-boundary or all-layout qualification. |
+| Uniform restart | Ten SRT/MRT checkpoint GPU tests passed on `8403956`. | The new representative TRT same-method double-buffer continuation test still requires its own GPU receipt; no LES, Jet or refined checkpoint claim follows. |
+| Taylor–Green refinement | Independent topology checks for the exact TG setup are available. The companion phase regression checks the actual lattice-rate configuration. | The physical interface-accuracy gate remains open. Topology and phase setup checks do not establish convergence or close that gate. |
+| Porous Couette and annulus | Prior campaign evidence exists and must remain associated with its original setup and source. | The reconstructed annulus is a validation helper, not qualification of a public boundary-condition API. Do not infer absence of evidence from a missing current rerun. |
+| LES and other combinations | Separate scalar and uniform-flow normalization evidence exists for the reviewed temporary correction. | That correction is not integrated into this source checkpoint. Do not transfer its acceptance to unchanged production, or to TRT/MRT LES, forcing, EsoPull, objects or refinement. |
+
+Historical uniform SRT cylinder-startup and inlet-rounding findings remain separate
+from the passed MRT DFG result. Raw exchange/window consistency is a measurement
+contract, not a general stress or force-accuracy certification. Existing source-pinned
+runners may intentionally reject a newer checkout; never silently repin old receipts.
+
+## Small next batch
+
+1. Cross-review the added TRT checkpoint and actual-rate TG phase tests, then run
+   their applicable checks with exact checkout/import receipts. Preserve the
+   existing SRT checkpoint cases and numerical criteria.
+2. Run the single added TRT checkpoint GPU case with the same method on both sides
+   of save/load. Record its result separately from the ten existing SRT/MRT passes.
+3. Continue the existing TG accuracy investigation using its actual fine-level
+   rates and the established analytic convention. Keep the accuracy gate open
+   until the relevant physical comparison passes.
+
+No repeat of the accepted full MRT DFG or Cartesian expansion across unsupported
+combinations is required by this document. Main integration remains a separate
+review of complete, bounded feature scope.