groundhog-hpc 0.8.0

Iterative HPC function development. As many 'first tries' as you need.

Home of hog ☀️🦫🕳️

Quickstart

Groundhog makes it easy to run, tweak, and re-run python functions on HPC clusters via Globus Compute using simple decorators.

Groundhog automatically manages remote environments (powered by uv)—just update Python versions or dependencies in your script, no SSH needed.

Key concepts:

• @hog.function() - Configures a function to run on a Globus Compute endpoint. Decorator kwargs (like endpoint, account) become the default user_endpoint_config.
• @hog.harness() - Marks a local entry point that orchestrates remote calls via .remote() or .submit(). Can also parse CLI arguments (example).
• The desired remote Python environment (version and dependencies) is specified alongside your code via PEP 723 metadata. You don't manage any remote state.

# /// script
# requires-python = ">=3.12,<3.13"
# dependencies = [
#     numpy,
# ]
#
# [tool.hog.tutorial]  # Globus Compute Tutorial Endpoint
# endpoint = "4b116d3c-1703-4f8f-9f6f-39921e5864df"
#
# ///

import groundhog_hpc as hog

@hog.function(endpoint='tutorial') # points to [tool.hog.tutorial] config
def compute(x: int) -> int:
    import numpy as np
    return int(np.sum(range(x)))

@hog.harness()
def main():
    result = compute.remote(100)
    print(result)

Run with: hog run myscript.py main

---

see also: examples