VoxelForge/Standalone/Research/WorldAssemblerM3/Linux86/RunL86CacheScaling.py

#!/usr/bin/env python3
"""Interleaved W8/W12 cold-readiness comparison for L86 exact passage result reuse."""
from __future__ import annotations

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

ROOT = Path(__file__).resolve().parents[4]
OUT = ROOT / "Saved/Linux/l86/cache-scaling"
DATA = ROOT / "Standalone/data/owner_world.json"
BINARY = ROOT / "Saved/Linux/l86/bin/VoxelForgeCoreHeadless-clang19"
WORKERS = (8, 12)
MODES = ("direct", "cache")
CAMERAS = {
    0: (8.0, 16.0, -372.2626953125),
    1: (25.8369140625, 3.19921875, -737.0),
    2: (19.166015625, -14.2001953125, -753.3740234375),
    3: (-20.0625, -5.068359375, -900.353515625),
}
EXPECTED_NEAR = {0: 1408, 1: 1216, 2: 1344, 3: 1344}
EXPECTED_WHOLE = {0: 2850, 1: 2249, 2: 2233, 3: 2329}
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,
)
CACHE = re.compile(
    r"^\[L86PassageResultCache\] hits=(\d+) misses=(\d+) entries=(\d+) "
    r"estimated_bytes=(\d+) exact_bake_ms=([0-9.]+) bake_mode=(\S+)$", re.M)


def run(strate: int, workers: int, mode: str, phase: str, replicate: int) -> dict:
    OUT.mkdir(parents=True, exist_ok=True)
    label = f"s{strate}-w{workers}-{mode}-{phase}{replicate}"
    log = OUT / f"{label}.log"
    camera = ",".join(f"{value:.12f}" for value in CAMERAS[strate]) + ",0,0"
    cache_value = "1" if mode == "cache" else "0"
    command = [
        "taskset", "-c", "0-11", "prlimit", "--as=6442450944", "--",
        "env", "VOXELFORGE_L81_PASSAGE_BOUND_PADDING_VOXELS=0",
        f"VOXELFORGE_L86_PASSAGE_RESULT_CACHE={cache_value}", "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",
    ]
    started = time.monotonic()
    with log.open("w", encoding="utf-8") as stream:
        result = subprocess.run(command, cwd=ROOT, stdout=stream,
                                stderr=subprocess.STDOUT, check=False, timeout=1740)
    output = log.read_text(encoding="utf-8", errors="replace")
    pipeline_match, near_match = PIPELINE.search(output), NEAR.search(output)
    if result.returncode or pipeline_match is None or near_match is None:
        raise RuntimeError(f"cache scaling run failed: exit={result.returncode} {log}")
    near_done, near_total = int(near_match.group(1)), int(near_match.group(2))
    if (near_done, near_total) != (EXPECTED_NEAR[strate], EXPECTED_NEAR[strate]):
        raise RuntimeError(f"incomplete near set in {log}: {near_done}/{near_total}")
    fields = dict(FIELDS.findall(pipeline_match.group(1)))
    whole_done, whole_total = (int(value) for value in fields["whole_tiles"].split("/"))
    if (whole_done, whole_total) != (EXPECTED_WHOLE[strate], EXPECTED_WHOLE[strate]):
        raise RuntimeError(f"incomplete whole set in {log}: {whole_done}/{whole_total}")
    cache_rows = CACHE.findall(output)
    if mode == "cache" and not cache_rows:
        raise RuntimeError(f"cache mode produced no cache statistics: {log}")
    cache_totals = {
        "hits": sum(int(row[0]) for row in cache_rows),
        "misses": sum(int(row[1]) for row in cache_rows),
        "entries": sum(int(row[2]) for row in cache_rows),
        "estimated_bytes": sum(int(row[3]) for row in cache_rows),
        "exact_bake_ms": sum(float(row[4]) for row in cache_rows),
        "bake_mode": sorted({row[5] for row in cache_rows}),
    }
    row = {
        "strate": strate, "workers": workers, "mode": mode,
        "phase": phase, "replicate": replicate,
        "exit_code": result.returncode, "wall_s": round(time.monotonic() - started, 3),
        "lod02_ready_s": float(near_match.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")),
        "near_tiles": near_total, "whole_tiles": whole_total,
        "cache": cache_totals, "log": str(log.relative_to(ROOT)),
    }
    print("[L86CacheScalingRun] " + json.dumps(row, sort_keys=True), flush=True)
    return row


def main() -> int:
    if not BINARY.is_file() or not DATA.is_file():
        raise SystemExit(f"missing input: {BINARY} or {DATA}")
    rows = []
    for strate in range(4):
        for workers in WORKERS:
            for mode in MODES:
                run(strate, workers, mode, "warmup", 0)
        worker_orders = ((8, 12), (12, 8), (8, 12))
        mode_orders = (("direct", "cache"), ("cache", "direct"),
                       ("direct", "cache"))
        for replicate, (worker_order, mode_order) in enumerate(
                zip(worker_orders, mode_orders), 1):
            for workers in worker_order:
                for mode in mode_order:
                    rows.append(run(strate, workers, mode, "measured", replicate))
                    (OUT / "runs-partial.json").write_text(
                        json.dumps(rows, indent=2) + "\n", encoding="utf-8")
    summary = []
    for strate in range(4):
        for workers in WORKERS:
            entry = {"strate": strate, "workers": workers, "measured_runs": 3}
            for mode in MODES:
                subset = [row for row in rows if row["strate"] == strate
                          and row["workers"] == workers and row["mode"] == mode]
                for metric in ("lod02_ready_s", "whole_view_ready_s", "shape_s", "far_shape_s"):
                    values = [row[metric] for row in subset]
                    entry[f"{mode}_{metric}"] = values
                    entry[f"{mode}_{metric}_median"] = statistics.median(values)
                    entry[f"{mode}_{metric}_min"] = min(values)
                    entry[f"{mode}_{metric}_max"] = max(values)
                entry[f"{mode}_cache"] = [row["cache"] for row in subset]
            summary.append(entry)
    (OUT / "summary.json").write_text(
        json.dumps({"runs": rows, "summary": summary}, indent=2) + "\n",
        encoding="utf-8",
    )
    return 0


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