Microbench

Microbench captures runtime and environment metadata for Python code to support performance benchmarking and reproducible computational analyses.

Key Features:

• Automated Metadata Capture: Automatically records execution time, software package versions, environment variables, hardware information (CPU and RAM usage), and Python version.
• Extensibility via plugins: Provides a plugin system to add custom metadata collection functionality.
• Reproducibility Support: Logs package versions and hardware specifications to aid replication of analyses across different systems and over time.
• Local and centralized storage: Stores captured metadata in a structured format to local files or a Redis database.

Scientific Applications:

• Computational systems biology: Addresses performance variability and reproducibility challenges across heterogeneous compute environments.
• Benchmarking and reproducibility analysis: Enables benchmarking of code execution and examination of environment metadata for reproducibility investigations.

Methodology:

Automatically capture comprehensive metadata during execution of Python functions via plugins and store the captured metadata in a structured format locally or in a Redis database.