---
id: ppft
version: "1.7.8"
license: BSD-3-Clause
license_treatment: permissive
maintenance: active
---

# ppft — distributed and parallel Python

License: permissive · Maintenance: active · Downloads: 17.9M/mo

## What it is and what it does

ppft is a fork of Parallel Python that brings the original package into modern Python with pip and setuptools support and enhanced serialization. It provides a job-server model for parallel execution: you create a Server, submit functions with their arguments and dependencies, and retrieve results asynchronously. Internally, ppft uses separate processes and inter-process communication to work around Python's Global Interpreter Lock, enabling true parallelism on multi-core systems and across networked computers.

The package handles the complexity of process management, load balancing, and fault tolerance automatically. Jobs can run locally on detected processor cores, on remote nodes via ppserver daemons, or both with dynamic load balancing. It supports function serialization by source code extraction, automatic processor detection, SHA-based network authentication, and statistics reporting on job execution.

Use it for:

- Distribute CPU-intensive computations across multiple cores on a single machine without threading bottlenecks.
- Execute long-running tasks on remote cluster nodes and retrieve results from a local client.
- Run data analysis or scientific calculations in parallel with automatic load balancing across heterogeneous hardware.
- Scale batch processing jobs across a network of computers with fault tolerance and dynamic resource allocation.
- Parallelize embarrassingly parallel workloads with minimal code changes.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

ppft executes Python functions in parallel across multiple processors or networked computers, bypassing the GIL by using separate processes and inter-process communication.

Yes, if you need straightforward job-based parallelism for CPU-bound Python code across multiple cores or machines. The low install friction, active maintenance, permissive license, and no runtime dependencies make it a practical choice. Consider it especially if you want to avoid the complexity of other frameworks or if you need both local and remote execution in one model. Not necessary if you're already using other multiprocessing libraries that fit your use case.

## Install

pip install ppft
uv add ppft
poetry add ppft

## Installing ppft

Before you install: Low friction installation via pip with no runtime dependencies. Actively maintained with recent commits and a stable release history since 2015.

License in practice: BSD-3-Clause permissive license allows commercial and private use with minimal restrictions; you must include a copy of the license and disclaimer.

Quickstart:

pip install ppft

import ppft as pp
import math

job_server = pp.Server()
f1 = job_server.submit(math.sin, (math.pi/2,), (), ('math',))
result = f1()
job_server.print_stats()

Requires Python >=3.9; if Parallel Python is already installed, uninstall it first to avoid import conflicts.

Verify before relying:

- Whether enhanced serialization using dill.source is automatic or requires explicit opt-in beyond installing ppft[dill]
- Performance characteristics and overhead compared to other multiprocessing frameworks for typical workloads
- Network security details beyond SHA-based authentication mentioned in the description

## Package facts

- License: BSD-3-Clause (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 17.9M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

parallel python execution, distributed computing python, multiprocessing job server, python cluster computing, bypass GIL parallelization, remote job execution, SMP parallel processing, parallel-computing, distributed-execution, cluster-computing

[View on SkillFed](https://skillfed.io/packages/ppft) · [View on PyPI](https://pypi.org/project/ppft/)