VoxelForge/Standalone/Research/WorldAssemblerM3/Linux140/RunL140Timing.py

#!/usr/bin/env python3
"""L80 timing: native L140 vs L139 HEAD and assembler/world, all owner strata."""
from __future__ import annotations

import json
import os
import re
import statistics
import subprocess
import time
from pathlib import Path

ROOT = Path(__file__).resolve().parents[4]
OUT = ROOT / "Saved/Linux/l140/timing"
DATA = ROOT / "Standalone/data/owner_world.json"
BINARIES = {
    "assembler-world": (
        ROOT / "Saved/Linux/l137/bin/VoxelForgeCoreHeadless-l137-aa-a",
        "4499fb2a8a4ede7d6cbc68d1b794f104658cdaae93ed9cacbc91fe7ff3fc3621"),
    "head": (
        ROOT / "Saved/Linux/l139/bin/VoxelForgeCoreHeadless-l139-clang",
        "b8603b6cb3975ca0f9669759e4e2464561c557fdb7aa6836acef123ecf0b99d1"),
    "l140-native-fi": (
        ROOT / "Saved/Linux/l140/bin/VoxelForgeCoreHeadless-l140-clang",
        "6205e6c180ef32c41792dc3e0d8bf443a96df41974f40334af4bf3a157ab88da"),
}
CAMERAS = {
    # The camera override also selects the owner SurfaceWorld departure mouth.
    # Hold XY to the seed-0 native-FI valid mouth so every layer/version shares one layout.
    0: (25.8369140625, 3.19921875, -372.2626953125),
    1: (25.8369140625, 3.19921875, -737.0),
    2: (25.8369140625, 3.19921875, -753.3740234375),
    3: (25.8369140625, 3.19921875, -900.353515625),
}
PIPELINE_RE = re.compile(r"^\[StreamPipeline\] (.*)$", re.M)
FIELDS_RE = re.compile(r"([a-zA-Z0-9_]+)=([^ ]+)")
NEAR_RE = re.compile(
    r"^\[WorldAssemblerNearLODReady\] lods=0-2 tiles=(\d+)/(\d+) cold_ready_s=([0-9.]+)$",
    re.M)
STAMP_RE = re.compile(r"^\[VoxelForgeSourceStamp\] sha256=([0-9a-f]{64})$", re.M)
L140_AA_FLOOR_PCT = 1.0  # L137's same-binary A/A median control.


def quiet_machine() -> tuple[float, float]:
    load_1m = os.getloadavg()[0]

    def totals() -> tuple[int, int]:
        fields = Path("/proc/stat").read_text().splitlines()[0].split()
        values = [int(value) for value in fields[1:]]
        return sum(values), values[3] + values[4]

    before = totals()
    time.sleep(2.0)
    after = totals()
    idle = 100.0 * (after[1] - before[1]) / max(1, after[0] - before[0])
    return load_1m, idle


def run(strate: int, workers: int, version: str, phase: str, replicate: int) -> dict[str, object]:
    binary, expected_stamp = BINARIES[version]
    camera = ",".join(f"{value:.12f}" for value in CAMERAS[strate]) + ",0,0"
    label = f"s{strate}-w{workers}-{version}-{phase}{replicate}"
    log = OUT / f"{label}.log"
    command = ["taskset", "-c", "0-11", "prlimit", "--as=6442450944", "--",
        "timeout", "1740s", 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",
        "--recipe=surface-world"]
    if version == "l140-native-fi":
        command.append("--native-owner-fi")
    started = time.monotonic()
    with log.open("w", encoding="utf-8") as output:
        proc = subprocess.run(command, cwd=ROOT, stdout=output,
            stderr=subprocess.STDOUT, check=False, timeout=1740)
    content = log.read_text(encoding="utf-8", errors="replace")
    pipeline, near, stamp = PIPELINE_RE.search(content), NEAR_RE.search(content), STAMP_RE.search(content)
    if (proc.returncode or pipeline is None or near is None or stamp is None
            or stamp.group(1) != expected_stamp):
        raise SystemExit(f"L80 run failed or stamp mismatch: {log}; exit={proc.returncode}")
    if version == "l140-native-fi" and not re.search(
            r"^\[FloatingIslandWorldAssembly\].*dispatch=FloatingIslands-native .*status=PASS$",
            content, re.M):
        raise SystemExit(f"candidate timing did not bind native FloatingIslands: {log}")
    fields = dict(FIELDS_RE.findall(pipeline.group(1)))
    row = {"strate": strate, "workers": workers, "version": version,
        "phase": phase, "replicate": replicate, "source_stamp": stamp.group(1),
        "lod02_ready_s": float(near.group(3)),
        "whole_view_ready_s": float(fields["whole_view_ready"].removesuffix("s")),
        "whole_tiles": int(fields["whole_tiles"].split("/")[0]),
        "wall_s": round(time.monotonic() - started, 3), "log": str(log.relative_to(ROOT))}
    print("[L140TimingRun] " + json.dumps(row, sort_keys=True), flush=True)
    return row


def main() -> int:
    if any(not path.is_file() for path, _ in BINARIES.values()) or not DATA.is_file():
        raise SystemExit("missing L80 timing binaries or owner fixture")
    OUT.mkdir(parents=True, exist_ok=True)
    load_1m, idle = quiet_machine()
    (OUT / "machine-load-before.txt").write_text(
        f"1-minute load: {load_1m:.2f}\n2-second CPU idle: {idle:.1f}%\n")
    print(f"[L140TimingGate] load_1m={load_1m:.2f} cpu_idle_2s={idle:.1f}%", flush=True)
    if load_1m >= 1.5 or idle < 97.0:
        raise SystemExit("timing refused: machine is above the L80 quiet-load gate")
    rows: list[dict[str, object]] = []
    for strate in range(4):
        for workers in (8, 12):
            for version in BINARIES:
                run(strate, workers, version, "warmup", 0)
            for replicate in range(1, 6):
                order = tuple(BINARIES)
                order = order[replicate % len(order):] + order[:replicate % len(order)]
                if replicate % 2 == 0:
                    order = tuple(reversed(order))
                for version in order:
                    rows.append(run(strate, workers, version, "measured", replicate))
                    (OUT / "runs-partial.json").write_text(json.dumps(rows, indent=2) + "\n")
    measured = [row for row in rows if row["phase"] == "measured"]
    summary = []
    for strate in range(4):
        for workers in (8, 12):
            group = [row for row in measured if row["strate"] == strate and row["workers"] == workers]
            medians = {version: {key: statistics.median(
                row[key] for row in group if row["version"] == version)
                for key in ("lod02_ready_s", "whole_view_ready_s")} for version in BINARIES}
            entry: dict[str, object] = {"strate": strate, "workers": workers,
                "replicates_per_version": 5, "medians": medians}
            for key in ("lod02_ready_s", "whole_view_ready_s"):
                for baseline in ("head", "assembler-world"):
                    entry[f"l140_vs_{baseline}_{key}_delta_pct"] = round(
                        100.0 * (medians["l140-native-fi"][key] / medians[baseline][key] - 1.0), 2)
                    entry[f"l140_vs_{baseline}_{key}_paired_deltas_pct"] = [round(100.0 * (
                        next(row[key] for row in group if row["version"] == "l140-native-fi"
                             and row["replicate"] == replicate)
                        / next(row[key] for row in group if row["version"] == baseline
                               and row["replicate"] == replicate) - 1.0), 3)
                        for replicate in range(1, 6)]
            summary.append(entry)
    result = {"a_a_noise_floor_pct": L140_AA_FLOOR_PCT,
        "a_a_reference": "L137 same-binary median control; assembler/world source stamp",
        "timing_protocol": "quiet machine; one warm-up per binary per cell; five interleaved measured runs; W8/W12; all four owner layers",
        "machine": {"load_1m": load_1m, "cpu_idle_2s_pct": idle},
        "source_stamps": {name: stamp for name, (_path, stamp) in BINARIES.items()},
        "runs": rows, "summary": summary}
    (OUT / "summary.json").write_text(json.dumps(result, indent=2) + "\n")
    print(f"[L140TimingSummary] measured_runs={len(measured)} warmups=24 aa_floor_pct=1.0 status=PASS",
          flush=True)
    return 0


if __name__ == "__main__":
    raise SystemExit(main())