Skip to content

Architectural: Implement Support for Concurrent Experiments #2

@nverchev

Description

@nverchev

Currently, the package's design relies on shared global state, which only allows a single experiment to run at a time. This prevents users from running multiple experiments concurrently.

FUTURE GOAL: Support for Python's threading and multiprocessing capabilities.

TODO:

  • make core components thread-safe (e.g., using threading.local for per-thread data, or locks for shared resources)
  • ensure experiments running in different processes cannot have the same run name.
  • stress and concurrency tests to verify that there is no leakage between experimental runs or interference in logging.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions