# uqfoundation/pathos

parallel graph management and execution in heterogeneous computing

Repository: https://github.com/uqfoundation/pathos
Canonical: https://ross.abutalabs.com/products/pathos
Homepage: http://pathos.rtfd.io
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-06-22T01:58:48+00:00

## Health v2 (maintenance only)
Score: 72/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 88, release rhythm 34, longevity 100
- inputs: {"age_days": 4814, "days_push": 73, "days_rel": 226, "gap_med": 237.5, "n_releases_24m": 3}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1484, forks 89 (observed 2026-08-28T04:04:51.361940+00:00)

## What it is
pathos is a Python framework for parallel graph management and execution across heterogeneous computing resources. It provides drop-in parallel and distributed versions of map and apply, with pluggable launchers including ssh-based, MPI-based, and multiprocessing backends.

## Use cases
- parallelize a serial Python loop with minimal refactoring
- run map/apply across multiple machines over ssh
- launch MPI-based parallel jobs on a cluster
- nest parallel maps for hierarchical heterogeneous computing
- serialize Python functions and closures for remote execution
- monitor and control remote RPC services from Python

## When to choose
- you want drop-in parallel map without rewriting your codebase
- you need one codebase spanning serial, threaded, multiprocess, ssh, and MPI execution
- you need to pickle nearly any Python object (via dill) for distributed tasks
- you want hierarchical/nested parallelism across heterogeneous resources

## When to avoid
- you need fine-grained control over MPI internals or node-level HPC scheduling
- you need GPU/accelerator parallelism (use CUDA or Dask instead)
- you want a full distributed data pipeline framework like Dask or Ray
- you need strict async/await concurrency rather than process-based parallelism

## Facets
- artifact type: library
- maturity: active
- function: concurrency, serialization, caching, rpc
- domain: microservices, developer-tools
- platform: python, cross-platform, cli
- tags: parallel-computing, parallel-map, heterogeneous-computing, ssh-launcher, mpi, multiprocessing, dill, automation

## Member repositories
- uqfoundation/pathos (main) score 72

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:51.361940+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:34:01.422364+00:00, confidence not recorded.
  - readme: https://github.com/uqfoundation/pathos (fetched 2026-08-28T04:04:51.361940+00:00, sha 1729357e87dd)
  - registry_pypi: https://pypi.org/pypi/pathos/json (fetched 2026-08-29T11:40:59.622587+00:00, sha 65f16286ef2f)
- Data as of 2026-08-30T08:39:29.467469+00:00.