Skip to content

bench_brain.py reference

Spatial brain performance benchmarking utility.

Description

The bench_brain.py tool measures read latency and byte throughput across brain assets and custom skills, calibrating performance multipliers for native OS and simulated mobile FUSE environments.

Script path

tools/bench_brain.py

CLI signature

uv run python tools/bench_brain.py

Inputs

Scans two target paths: - .agents/brain - .agents/skills

Outputs

Telemetry results printed directly to the shell terminal: - Count of scanned files and total parsed byte volumes. - Total read times for native OS (in milliseconds). - Estimated mobile (Samsung Note 10) Termux FUSE latency extrapolations (using a default 3.5x multiplier).

Dependencies

  • Python: Standard library only (no external packages required).

Internal Python API

get_files(directories, extension)

Recursively gathers matching files. - Arguments: directories (string list), extension (string, defaults to .md). - Returns: String path list.

bench_read(files)

Evaluates overall reading performance. - Returns: Tuple containing (total_bytes, elapsed_seconds).

Deep State of Mind (DSOM) For My AI Protocol | Harisfazillah Jamel (LinuxMalaysia) | 2026-08-14 Standard: UK English | DBP-standard Bahasa Melayu Malaysia (Piawai) | GNU General Public License v3.0