config.ADAPTIVE_CELLS (dialog checkbox "adaptive cells", off by
default; standalone --adaptive) routes run_solve through
fill_resistance/adaptive.py: per-layer balanced leaf grids where every
non-uniform fine cell (electrodes, 1D chain cells, buildup, via-mouth
thickness map) is pinned at the fine size, leaf faces via the
series-half-cell rule, chain links and barrels re-attached by node id,
connectivity restriction and both contact models on the leaf graph via
solver cores extracted for reuse (_equipotential_core, _uniform_core,
_conductance_params, _barrel_links). All fields (V, |J|, power density)
are computed per leaf and expanded to the fine grid, so plots, summary
and dumps are unchanged.
Element sizes: minimum = the grid cell size itself (auto / dialog /
CELL_UM_OVERRIDE); maximum = ADAPTIVE_MAX_CELL_UM (2 mm default);
ADAPTIVE_GUARD sets the clearance a block needs to grow.
Measured end-to-end (feature-dense 120x120 plate, h=50um): 25.9 s ->
5.4 s, 5.58M -> 823k unknowns, R -1.1%. Accuracy documented honestly:
coarse-fine interfaces carry a first-order tangential flux error
biasing R low by ~0.5-2% depending on geometry (worst on narrow
strips); the earlier assumption that linear fields solve exactly on the
leaf graph was wrong - offset centers across size transitions leave an
unpaired residue. Gradient-corrected interface fluxes remain as phase 4
if tighter accuracy per leaf is needed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
fill_resistance/quadtree.py decomposes a layer's fine copper mask into
2:1-balanced power-of-two leaves: boundary and keep-fine cells stay at
the fine size, interiors coarsen with their Chebyshev distance to the
nearest feature (guard factor, default 4), and an explicit enforcement
pass splits any leaf more than twice an edge-adjacent neighbor. Face
conductances use the series-half-cell rule, which reduces to the
production harmonic mean for equal sizes and EXACTLY to sigma in the
uniform limit - verified edge-for-edge against solver.build_edges and
to rel 1e-12 in R against run_solve, so the exact-value test suite
stays authoritative for this engine.
Measured (feature-dense 120x120 plate, h=50um, production AMG solver):
uniform 5.58M unknowns ~35s; adaptive guard=4 823k / ~8s at -1.1%;
guard=8 1.74M / ~12s at -0.47%. tools/adaptive_proto.py now benchmarks
the engine itself.
Not yet wired into the pipeline: phase 2 ports electrodes, barrels,
1D chains, buildup and field output onto leaves.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
janik