pdc_benchmark.py

"""PDC benchmarks via compiled C backend (Sequential, OpenMP, POSIX Threads)."""

import subprocess
import sys
from pathlib import Path

CHART_CATEGORIES = [
    "Sequential",
    "OpenMP (2)",
    "OpenMP (4)",
    "POSIX Threads (4)",
]

APP_DIR = Path(__file__).resolve().parent
BACKEND_DIR = APP_DIR / "c_backend"
PROJECT_ROOT = APP_DIR.parent


def _bench_binary() -> Path:
    name = "search_bench.exe" if sys.platform == "win32" else "search_bench"
    for base in (BACKEND_DIR, PROJECT_ROOT):
        candidate = base / name
        if candidate.exists():
            return candidate
    return BACKEND_DIR / name


def _bench_cwd(bench: Path) -> Path:
    return bench.parent


def _label_for(method: str, workers: int) -> str:
    if method == "Sequential":
        return "Sequential"
    if method == "OpenMP":
        return f"OpenMP ({workers})"
    if method == "POSIX Threads":
        return f"POSIX Threads ({workers})"
    return method


def run_backend_benchmark(method_type: str, worker_count: int, keyword: str) -> float:
    bench = _bench_binary()
    if not bench.exists():
        raise FileNotFoundError(_build_help())
    proc = subprocess.run(
        [str(bench), method_type, str(worker_count), keyword],
        capture_output=True,
        text=True,
        cwd=str(_bench_cwd(bench)),
    )
    if proc.returncode != 0:
        raise RuntimeError(proc.stderr.strip() or "C benchmark failed")
    return float(proc.stdout.strip())


def _build_help() -> str:
    return (
        "Compiled benchmark not found.\n"
        "Build locally:\n"
        "  cd c_backend && gcc -O2 -fopenmp -pthread search_bench.c hostel_mock_data.c "
        "-o search_bench -lm"
    )


def _parse_row(line: str) -> dict:
    parts = line.split()
    if parts[0] == "POSIX" and len(parts) > 1 and parts[1] == "Threads":
        method = "POSIX Threads"
        workers = int(parts[2])
        fee_t, search_t, total_t = float(parts[3]), float(parts[4]), float(parts[5])
        found, speedup = int(parts[6]), float(parts[7])
    else:
        method = parts[0]
        workers = int(parts[1])
        fee_t, search_t, total_t = float(parts[2]), float(parts[3]), float(parts[4])
        found, speedup = int(parts[5]), float(parts[6])

    return {
        "Method": method,
        "Workers": workers,
        "Fee (s)": fee_t,
        "Search (s)": search_t,
        "Time (s)": total_t,
        "Found": found,
        "Speedup": speedup,
        "Label": _label_for(method, workers),
    }


def run_all_benchmarks(students, keyword="Ali"):
    del students
    bench = _bench_binary()
    if not bench.exists():
        raise FileNotFoundError(_build_help())

    proc = subprocess.run(
        [str(bench), "all", keyword],
        capture_output=True,
        text=True,
        cwd=str(_bench_cwd(bench)),
    )
    if proc.returncode != 0:
        raise RuntimeError(proc.stderr.strip() or "C benchmark failed")

    rows = [_parse_row(line) for line in proc.stdout.strip().splitlines() if line.strip()]
    if not rows:
        raise RuntimeError("No benchmark results from C backend")
    return rows, keyword