Spaces:
Running
Running
File size: 5,699 Bytes
31226fd | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 | """Run a bounded, reproducible MHRN scaling benchmark ladder.
Examples:
python scripts/benchmark_ladder.py --tiers 100,500,5000 --ticks 20
python scripts/benchmark_ladder.py --tiers 5000,50000,1000000 --ticks 10 --allow-large
The benchmark reports construction, Python allocation peak, tick cost and
throughput. It does not claim scientific performance evidence; hardware,
Python version and configuration are recorded in the JSON output.
"""
from __future__ import annotations
import argparse
import json
import math
import platform
import random
import sys
import time
import tracemalloc
from pathlib import Path
from typing import Any
REPO_ROOT = Path(__file__).resolve().parent.parent
if str(REPO_ROOT) not in sys.path:
sys.path.insert(0, str(REPO_ROOT))
DEFAULT_TIERS = (5_000, 25_000, 100_000, 1_000_000)
LARGE_TIER_LIMIT = 100_000
def build_network(neuron_count: int, seed: int, connections_per_neuron: int = 0) -> Any:
from src.core.network import Brain5DConfig, NeuralNetwork
from src.core.spatial_index import linear_to_5d
side = max(1, math.ceil(neuron_count ** (1.0 / 5.0)))
config = Brain5DConfig.from_dict(
{
"dimensions": [side, side, side, side, side],
"network": {
"initial_connections_per_neuron": connections_per_neuron,
"neighbour_radius": 1.0,
},
}
)
network = NeuralNetwork(config, random.Random(seed))
for index in range(neuron_count):
network.add_neuron(linear_to_5d(index, config.dimensions))
if connections_per_neuron:
network.initialize_random_connections(
connections_per_neuron=connections_per_neuron,
radius=1.0,
)
return network
def run_tier(
neuron_count: int,
ticks: int,
seed: int,
connections_per_neuron: int = 0,
) -> dict[str, Any]:
"""Run one bounded tier and return engineering-only performance metrics."""
tracemalloc.start()
try:
started = time.perf_counter()
network = build_network(neuron_count, seed, connections_per_neuron)
construction_seconds = time.perf_counter() - started
step_started = time.perf_counter()
for _ in range(ticks):
network.step()
step_seconds = time.perf_counter() - step_started
_, peak_bytes = tracemalloc.get_traced_memory()
finally:
tracemalloc.stop()
mean_tick_cost_ms = (step_seconds * 1000.0 / ticks) if ticks else None
return {
"neurons": neuron_count,
"synapses": network.get_state_summary().get("synapses", 0),
"ticks": ticks,
"connections_per_neuron_requested": connections_per_neuron,
"construction_seconds": round(construction_seconds, 6),
"step_seconds": round(step_seconds, 6),
"mean_tick_cost_ms": (
round(mean_tick_cost_ms, 6) if mean_tick_cost_ms is not None else None
),
"python_peak_memory_bytes": int(peak_bytes),
"python_peak_bytes_per_neuron": round(peak_bytes / neuron_count, 3),
"ticks_per_second": round(ticks / step_seconds, 3) if step_seconds else None,
"neurons_per_second": (
round(neuron_count * ticks / step_seconds, 3) if step_seconds else None
),
"synapses_per_second": (
round(network.synapse_count * ticks / step_seconds, 3)
if step_seconds
else None
),
}
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument(
"--tiers",
default=",".join(str(value) for value in DEFAULT_TIERS),
help="Comma-separated neuron counts (default: 5000,25000,100000,1000000)",
)
parser.add_argument("--ticks", type=int, default=20)
parser.add_argument("--seed", type=int, default=42)
parser.add_argument("--connections-per-neuron", type=int, default=0)
parser.add_argument("--output", type=Path)
parser.add_argument(
"--allow-large",
action="store_true",
help="Allow tiers above the 100k safety limit.",
)
return parser.parse_args()
def main() -> int:
args = parse_args()
if args.ticks <= 0:
raise SystemExit("--ticks must be positive")
if args.connections_per_neuron < 0:
raise SystemExit("--connections-per-neuron must be non-negative")
tiers = tuple(sorted({int(value) for value in args.tiers.split(",") if value}))
if not tiers or any(value <= 0 for value in tiers):
raise SystemExit("--tiers must contain positive integers")
if not args.allow_large and any(value > LARGE_TIER_LIMIT for value in tiers):
raise SystemExit(
"tiers above 100000 require --allow-large; this prevents accidental 1M runs"
)
report = {
"schema_version": 2,
"benchmark": "mhrn_scaling_ladder",
"python": platform.python_version(),
"platform": platform.platform(),
"seed": args.seed,
"requested_ticks": args.ticks,
"connections_per_neuron": args.connections_per_neuron,
"memory_metric": "tracemalloc_python_peak_bytes",
"tiers": [
run_tier(
value,
args.ticks,
args.seed,
args.connections_per_neuron,
)
for value in tiers
],
"scientific_claim": False,
}
payload = json.dumps(report, indent=2)
if args.output:
args.output.parent.mkdir(parents=True, exist_ok=True)
args.output.write_text(payload + "\n", encoding="utf-8")
print(payload)
return 0
if __name__ == "__main__":
sys.exit(main())
|