core: outside-influence fast path for swept samples (L138, output = v236)

Column-cache census matches assembler/world (the L77 share rise was relative, not more calls). Reuses the TN
predicate and skips relief/mouth/shell work for stamped samples outside the 1.053 m influence. vs assembler/world:
whole view -0.5..-4.0% on all layers; near-ready Surface World +0.3/+1.7, Floating Islands -4.3/+2.9, Tunnel Network
-2.5/+2.9, Vertical Shafts +4.7/+3.3 (A/A ±1%). 16/16 seeds, fast tier, Clang/GCC green.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016dhxM43dnWvUu2B5Ph1iG1

Standalone/Research/WorldAssemblerM3/Linux138/AnalyzeL138PrechangeCacheCounts.py

@@ -0,0 +1,103 @@
+#!/usr/bin/env python3
+"""Summarize per-tile cache counters from the five L137 paired timing runs."""
+from __future__ import annotations
+
+import json
+import re
+import statistics
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[4]
+SOURCE = ROOT / "Saved/Linux/l137/timing/final-design1"
+OUT = ROOT / "Saved/Linux/l138/prechange/column-cache-counts.json"
+STAMPS = {
+    "assembler-world": "4499fb2a8a4ede7d6cbc68d1b794f104658cdaae93ed9cacbc91fe7ff3fc3621",
+    "l137": "7913d019d968de2dbaea2282b90fa8ff41b53f996c022e9e6fed92712cb40ecf",
+}
+FIELDS = re.compile(r"([a-zA-Z0-9_]+)=([^ ]+)")
+METRICS = (
+    "column_point_calls", "column_local_hits", "column_local_misses",
+    "column_conflict_misses", "column_shared_hits", "column_formula_evaluations",
+)
+
+
+def read_run(strate: int, workers: int, version: str, replicate: int) -> dict:
+    path = SOURCE / f"s{strate}-w{workers}-{version}-measured{replicate}.log"
+    if not path.is_file():
+        raise SystemExit(f"missing pre-change L137 run: {path}")
+    text = path.read_text(encoding="utf-8", errors="replace")
+    stamp = re.search(r"^\[VoxelForgeSourceStamp\] sha256=([0-9a-f]+)$", text, re.M)
+    if stamp is None or stamp.group(1) != STAMPS[version]:
+        raise SystemExit(f"unexpected source stamp in {path}")
+    rows = {}
+    for line in text.splitlines():
+        if not line.startswith("[WorldAssemblerLODCostDetail] "):
+            continue
+        fields = dict(FIELDS.findall(line))
+        rows[int(fields["lod"])] = fields
+    if set(rows) != set(range(9)):
+        raise SystemExit(f"incomplete LOD counters in {path}: {sorted(rows)}")
+    return {"path": str(path.relative_to(ROOT)), "rows": rows}
+
+
+def summarize(runs: list[dict], lods: range) -> dict:
+    per_run = []
+    for run in runs:
+        selected = [run["rows"][lod] for lod in lods]
+        tile_count = sum(int(row["tiles"]) for row in selected)
+        per_run.append({
+            "tiles": tile_count,
+            **{metric: sum(int(row[metric]) for row in selected) / tile_count
+               for metric in METRICS},
+        })
+    return {
+        "tiles_per_run": statistics.mean(row["tiles"] for row in per_run),
+        "mean_per_tile": {
+            metric: statistics.mean(row[metric] for row in per_run)
+            for metric in METRICS
+        },
+        "run_range_per_tile": {
+            metric: [min(row[metric] for row in per_run),
+                     max(row[metric] for row in per_run)]
+            for metric in METRICS
+        },
+    }
+
+
+def main() -> int:
+    result = {
+        "source": "five measured L137 L80 runs per version, worker count, and stratum",
+        "pre_change_stamps": STAMPS,
+        "cells": [],
+    }
+    for strate in range(4):
+        for workers in (8, 12):
+            versions = {
+                version: [read_run(strate, workers, version, replicate)
+                          for replicate in range(1, 6)]
+                for version in STAMPS
+            }
+            item = {"strate": strate, "workers": workers}
+            for window, lods in (("near_lod0_2", range(3)), ("whole_lod0_8", range(9))):
+                item[window] = {
+                    version: summarize(runs, lods)
+                    for version, runs in versions.items()
+                }
+                item[window]["candidate_minus_baseline_per_tile"] = {
+                    metric: (item[window]["l137"]["mean_per_tile"][metric]
+                             - item[window]["assembler-world"]["mean_per_tile"][metric])
+                    for metric in METRICS
+                }
+            item["logs"] = {
+                version: [run["path"] for run in runs]
+                for version, runs in versions.items()
+            }
+            result["cells"].append(item)
+    OUT.parent.mkdir(parents=True, exist_ok=True)
+    OUT.write_text(json.dumps(result, indent=2) + "\n", encoding="utf-8")
+    print(f"[L138PrechangeCacheCounts] cells={len(result['cells'])} output={OUT.relative_to(ROOT)}")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())