diag: passage vs boundary-band ceiling split (L88); coarse passage cache rejected
Verified ablations (zero passage work / zero band calls). LOD0-2 @W12: exact FI 2.27 / TN 2.25 s; passages off FI 0.81 / TN 0.85; bands off FI 2.11 / TN 2.06. Passages cost ~1.4-1.6 s of near readiness, bands ~0.1-0.2 s. A tile-local 4-voxel trilinear passage-delta cache increased carve segment visits (FI 34.5M -> 56.2M) and slowed LOD0-2 (FI 2.58 -> 3.77 s); removed. No world change. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016dhxM43dnWvUu2B5Ph1iG1
This commit is contained in:
Fr0zkaandClaude Opus 5.5 committed
1 parent
0fe7f371f7
commit
fa78192c86
5 files changed
+287
-17
No files matched your search
@@ -0,0 +1,35 @@
|
||||
# Linux 88 — passage and boundary cost
|
||||
|
||||
## Ceiling split
|
||||
|
||||
L80 protocol: one warmup, then three interleaved measured runs per mode, worker count and stratum. Values are median seconds to LOD0–2 ready / whole view ready.
|
||||
|
||||
| Stratum | Workers | Exact | Passages off, bands on | Bands off, passages on |
|
||||
|---|---:|---:|---:|---:|
|
||||
| FI | 8 | 2.419 / 3.696 | 0.808 / 1.738 | 2.306 / 3.556 |
|
||||
| FI | 12 | 2.273 / 3.463 | 0.812 / 1.683 | 2.114 / 3.290 |
|
||||
| TN | 8 | 2.458 / 3.624 | 0.912 / 1.814 | 2.222 / 3.308 |
|
||||
| TN | 12 | 2.253 / 3.294 | 0.851 / 1.683 | 2.056 / 3.061 |
|
||||
|
||||
Passage composition costs about 1.40–1.61 s of FI/TN LOD0–2 readiness. Boundary bands cost about 0.11–0.24 s. The split counters confirmed the ablations: passages-off had zero passage queries, iterations, carve/support visits and relief calls, while retaining 11.39M FI / 7.72M TN band calls in profiled tiles. Bands-off had zero band calls while passage iterations remained 14.06M FI / 12.43M TN. The profile covered 629 FI and 614 TN near tiles. Passage-off intentionally failed the route walkability check because the remaining bands close an authored mouth; a split-only diagnostic continued the timing run.
|
||||
|
||||
Exact-field readiness across all four strates:
|
||||
|
||||
| Stratum | W8 LOD0–2 / whole | W12 LOD0–2 / whole |
|
||||
|---|---:|---:|
|
||||
| SurfaceWorld | 1.093 / 2.340 | 1.024 / 2.157 |
|
||||
| FloatingIslands | 2.419 / 3.696 | 2.273 / 3.463 |
|
||||
| TunnelNetwork | 2.458 / 3.624 | 2.253 / 3.294 |
|
||||
| VerticalShafts | 1.312 / 2.386 | 1.128 / 2.080 |
|
||||
|
||||
## Prototype and decision
|
||||
|
||||
I tested a tile-local 4-voxel trilinear cache of passage-composition deltas, with exact evaluation near or across the zero surface. It did not remove enough route work: instrumented FI/TN runs recorded 4.75M / 7.71M exact coarse-node evaluations, and carve-segment visits increased from 34.5M / 33.1M to 56.2M / 76.5M.
|
||||
|
||||
Two unprofiled interleaved W8 pairs confirmed the regression. FI LOD0–2 median rose from 2.583 s exact to 3.765 s coarse; TN rose from 2.646 s to 4.649 s. Whole-view medians rose from 3.957 to 8.708 s for FI and 3.735 to 7.976 s for TN. The prototype was removed from source.
|
||||
|
||||
The pre-measurement acceptance bar was max displacement ≤5 cm, mean ≤1 cm per stratum, and zero sign flips on canonical walk-floor samples. Displacement was not measured. The prototype stopped at the speed gate, so post-prototype passage audits, captures, and Clang/GCC W1/W8/W12 comparisons were not run. No evaluator version bump or hash re-pins were made; version remains 214. The fast tier passed on the cleaned exact-field source: 274 checks in 311.3 s, stamp `81a934b80ed90d2ab0137345d15688a9bd201e4ec6ce7af94b08f4cb4e77c1a3`.
|
||||
|
||||
## Runs
|
||||
|
||||
All benchmark and verification jobs used CPUs 0–11; headless jobs were capped at 6 GiB. No run exceeded 30 minutes. The timing and per-tile work logs are under `Saved/Linux/l88/ceiling-split/` and `Saved/Linux/l88/coarse-prototype/`; the fast-tier log and summary are under `Saved/Linux/l88/logs/fast-tier.log` and `Saved/Linux/l88/verify-fast/fast-summary.json`. No Unreal `Source/` files changed, so no UE build was needed. No commit, push, or stash was made.
|
||||
@@ -0,0 +1,198 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Measure passage and boundary-band ceilings independently under the L80 protocol."""
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import json
|
||||
import re
|
||||
import statistics
|
||||
import subprocess
|
||||
import time
|
||||
from pathlib import Path
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[4]
|
||||
OUT = ROOT / "Saved/Linux/l88/ceiling-split"
|
||||
DATA = ROOT / "Standalone/data/owner_world.json"
|
||||
BINARY = ROOT / "Saved/Linux/l88/bin/VoxelForgeCoreHeadless-l88-clang19"
|
||||
STRATA = {
|
||||
"S0": {"id": 0, "camera": (8.0, 16.0, -372.2626953125), "near": 1408, "whole": 2850},
|
||||
"FI": {"id": 1, "camera": (25.8369140625, 3.19921875, -737.0), "near": 1216, "whole": 2249},
|
||||
"TN": {"id": 2, "camera": (19.166015625, -14.2001953125, -753.3740234375), "near": 1344, "whole": 2233},
|
||||
"S3": {"id": 3, "camera": (-20.0625, -5.068359375, -900.353515625), "near": 1344, "whole": 2329},
|
||||
}
|
||||
WORKERS = (8, 12)
|
||||
MODES = ("direct", "passages_off", "bands_off")
|
||||
PIPELINE = re.compile(r"^\[StreamPipeline\] (.*)$", re.M)
|
||||
FIELDS = re.compile(r"([a-zA-Z0-9_]+)=([^ ]+)")
|
||||
NEAR = re.compile(
|
||||
r"^\[WorldAssemblerNearLODReady\] lods=0-2 tiles=(\d+)/(\d+) cold_ready_s=([0-9.]+)$", re.M)
|
||||
PASSAGE_TILE = re.compile(r"^\[PassageComposerTile\] (.*)$", re.M)
|
||||
STAMP = re.compile(r"^\[VoxelForgeSourceStamp\] sha256=(\S+)$", re.M)
|
||||
|
||||
|
||||
def base_command(stratum: str, workers: int, mode: str, profile: bool = False) -> list[str]:
|
||||
spec = STRATA[stratum]
|
||||
camera = ",".join(f"{value:.12f}" for value in spec["camera"]) + ",0,0"
|
||||
child = [
|
||||
str(BINARY), str(DATA), "--world-assembler-prototype", "--stream-benchmark",
|
||||
"--stream-pipeline", f"--workers={workers}", "--clip-radius=3",
|
||||
"--max-clip-level=8", "--max-level=8", f"--camera={camera}",
|
||||
"--profile-lod-work", "--stream-scheduler-policy=3",
|
||||
]
|
||||
if profile:
|
||||
child.append("--profile-passage-composer")
|
||||
env = ["VOXELFORGE_L81_PASSAGE_BOUND_PADDING_VOXELS=0"]
|
||||
if mode == "passages_off":
|
||||
env.extend(("VOXELFORGE_L88_DISABLE_PASSAGES=1",
|
||||
"VOXELFORGE_L88_ALLOW_PASSAGE_ABLATION_WALKABILITY_FAILURE=1"))
|
||||
elif mode == "bands_off":
|
||||
env.append("VOXELFORGE_L88_DISABLE_BOUNDARY_BANDS=1")
|
||||
return [
|
||||
"taskset", "-c", "0-11", "prlimit", "--as=6442450944", "--", "env",
|
||||
*env, "timeout", "1740s", *child,
|
||||
]
|
||||
|
||||
|
||||
def execute(stratum: str, workers: int, mode: str, phase: str,
|
||||
replicate: int = 0, profile: bool = False) -> dict:
|
||||
OUT.mkdir(parents=True, exist_ok=True)
|
||||
label = f"{stratum.lower()}-w{workers}-{mode}-{phase}"
|
||||
if replicate:
|
||||
label += f"-r{replicate}"
|
||||
if profile:
|
||||
label += "-profile"
|
||||
log = OUT / f"{label}.log"
|
||||
started = time.monotonic()
|
||||
with log.open("w", encoding="utf-8") as stream:
|
||||
result = subprocess.run(base_command(stratum, workers, mode, profile), cwd=ROOT,
|
||||
stdout=stream, stderr=subprocess.STDOUT,
|
||||
check=False, timeout=1750)
|
||||
output = log.read_text(encoding="utf-8", errors="replace")
|
||||
pipeline, near, stamp = PIPELINE.search(output), NEAR.search(output), STAMP.search(output)
|
||||
if result.returncode or pipeline is None or near is None or stamp is None:
|
||||
raise RuntimeError(f"run failed or required output is missing: {log}; exit={result.returncode}")
|
||||
spec = STRATA[stratum]
|
||||
near_counts = (int(near.group(1)), int(near.group(2)))
|
||||
if near_counts != (spec["near"], spec["near"]):
|
||||
raise RuntimeError(f"incomplete near set in {log}: {near_counts}")
|
||||
fields = dict(FIELDS.findall(pipeline.group(1)))
|
||||
whole_counts = tuple(int(value) for value in fields["whole_tiles"].split("/"))
|
||||
if whole_counts != (spec["whole"], spec["whole"]):
|
||||
raise RuntimeError(f"incomplete whole view in {log}: {whole_counts}")
|
||||
ablation_walkability = "[L88PassageAblation]" in output
|
||||
if (mode == "passages_off") != ablation_walkability:
|
||||
raise RuntimeError(f"passage ablation walkability marker unexpected in {log}")
|
||||
row = {
|
||||
"stratum": stratum, "workers": workers, "mode": mode, "phase": phase,
|
||||
"replicate": replicate, "profile": profile, "exit_code": result.returncode,
|
||||
"wall_s": round(time.monotonic() - started, 3), "source_stamp": stamp.group(1),
|
||||
"lod02_ready_s": float(near.group(3)),
|
||||
"whole_view_ready_s": float(fields["whole_view_ready"].removesuffix("s")),
|
||||
"shape_s": float(fields["shape"].removesuffix("s")),
|
||||
"far_shape_s": float(fields["far_shape"].removesuffix("s")),
|
||||
"ablation_walkability_failed": ablation_walkability,
|
||||
"near_tiles": spec["near"], "whole_tiles": spec["whole"],
|
||||
"log": str(log.relative_to(ROOT)),
|
||||
}
|
||||
if profile:
|
||||
total = {name: 0 for name in (
|
||||
"boundary_band_queries", "queries", "passage_iterations",
|
||||
"carve_segment_visits", "support_segment_visits", "relief_field_calls",
|
||||
)}
|
||||
profiled_tiles = 0
|
||||
for match in PASSAGE_TILE.finditer(output):
|
||||
fields = dict(FIELDS.findall(match.group(1)))
|
||||
if int(fields["key"].rsplit("lod", 1)[1].rstrip(")")) > 2:
|
||||
continue
|
||||
profiled_tiles += 1
|
||||
for name in total:
|
||||
total[name] += int(fields[name])
|
||||
if not profiled_tiles:
|
||||
raise RuntimeError(f"passage/band profile rows missing in {log}")
|
||||
row.update({f"profile_{name}": value for name, value in total.items()})
|
||||
row["profiled_near_tiles"] = profiled_tiles
|
||||
print("[L88CeilingSplitRun] " + json.dumps(row, sort_keys=True), flush=True)
|
||||
return row
|
||||
|
||||
|
||||
def profile_modes() -> list[dict]:
|
||||
rows = [execute(stratum, 8, mode, "profile", profile=True)
|
||||
for stratum in ("FI", "TN") for mode in MODES]
|
||||
for stratum in ("FI", "TN"):
|
||||
subset = {row["mode"]: row for row in rows if row["stratum"] == stratum}
|
||||
direct, passages, bands = (subset[name] for name in MODES)
|
||||
if direct["profile_boundary_band_queries"] == 0 or direct["profile_queries"] == 0:
|
||||
raise RuntimeError(f"direct profile did not exercise both paths for {stratum}")
|
||||
if passages["profile_boundary_band_queries"] == 0 or any(
|
||||
passages[f"profile_{name}"] != 0
|
||||
for name in ("queries", "passage_iterations", "carve_segment_visits",
|
||||
"support_segment_visits", "relief_field_calls")
|
||||
):
|
||||
raise RuntimeError(f"passages-off profile did not isolate the band path for {stratum}")
|
||||
if bands["profile_boundary_band_queries"] != 0 or any(
|
||||
bands[f"profile_{name}"] == 0
|
||||
for name in ("queries", "passage_iterations", "carve_segment_visits")
|
||||
):
|
||||
raise RuntimeError(f"bands-off profile did not preserve passage work for {stratum}")
|
||||
(OUT / "profile-summary.json").write_text(json.dumps(rows, indent=2) + "\n", encoding="utf-8")
|
||||
return rows
|
||||
|
||||
|
||||
def timing_matrix() -> list[dict]:
|
||||
# One warmup for each complete stratum/worker/mode window.
|
||||
for stratum in STRATA:
|
||||
for workers in WORKERS:
|
||||
for mode in MODES:
|
||||
execute(stratum, workers, mode, "warmup")
|
||||
orders = (
|
||||
((8, "direct"), (8, "passages_off"), (8, "bands_off"),
|
||||
(12, "bands_off"), (12, "passages_off"), (12, "direct")),
|
||||
((12, "passages_off"), (12, "direct"), (12, "bands_off"),
|
||||
(8, "bands_off"), (8, "direct"), (8, "passages_off")),
|
||||
((8, "bands_off"), (8, "direct"), (8, "passages_off"),
|
||||
(12, "direct"), (12, "passages_off"), (12, "bands_off")),
|
||||
)
|
||||
measured = []
|
||||
for stratum in STRATA:
|
||||
for replicate, order in enumerate(orders, 1):
|
||||
for workers, mode in order:
|
||||
measured.append(execute(stratum, workers, mode, "measured", replicate))
|
||||
(OUT / "runs-partial.json").write_text(
|
||||
json.dumps(measured, indent=2) + "\n", encoding="utf-8")
|
||||
summary = []
|
||||
for stratum in STRATA:
|
||||
for workers in WORKERS:
|
||||
for mode in MODES:
|
||||
subset = [r for r in measured if r["stratum"] == stratum
|
||||
and r["workers"] == workers and r["mode"] == mode]
|
||||
if len(subset) != 3:
|
||||
raise RuntimeError(f"expected 3 measured runs for {stratum} W{workers} {mode}")
|
||||
item = {"stratum": stratum, "workers": workers, "mode": mode,
|
||||
"measured_runs": len(subset)}
|
||||
for metric in ("lod02_ready_s", "whole_view_ready_s", "shape_s", "far_shape_s"):
|
||||
values = [r[metric] for r in subset]
|
||||
item[metric] = values
|
||||
item[f"{metric}_median"] = statistics.median(values)
|
||||
summary.append(item)
|
||||
(OUT / "summary.json").write_text(
|
||||
json.dumps({"runs": measured, "summary": summary}, indent=2) + "\n",
|
||||
encoding="utf-8")
|
||||
return measured
|
||||
|
||||
|
||||
def main() -> int:
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument("--profiles-only", action="store_true")
|
||||
parser.add_argument("--timing-only", action="store_true")
|
||||
args = parser.parse_args()
|
||||
if not BINARY.is_file() or not DATA.is_file():
|
||||
raise SystemExit(f"missing input: {BINARY} or {DATA}")
|
||||
if not args.timing_only:
|
||||
profile_modes()
|
||||
if not args.profiles_only:
|
||||
timing_matrix()
|
||||
return 0
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
Reference in new issue
Block a user