Skip to main content

CPU parallelism for Trio

Project description

Welcome to trio-parallel!

CPU parallelism for Trio

License: Your choice of MIT or Apache License 2.0

Do you have CPU bound work that just keeps slowing down your event loop no matter what you try? Do you need to get all those cores humming at once? This is the library for you!

import multiprocessing
import trio
import trio_parallel
import time


def hard_work(n, x):
    t = time.perf_counter() + n
    y = x
    while time.perf_counter() < t:
        x = not x
    print(y, "transformed into", x)
    return x

async def too_slow():
    await trio_parallel.run_sync(hard_work, 20, False, cancellable=True)


async def amain():
    t0 = time.perf_counter()
    async with trio.open_nursery() as nursery:
        nursery.start_soon(trio_parallel.run_sync, hard_work, 3, True)
        nursery.start_soon(trio_parallel.run_sync, hard_work, 1, False)
        nursery.start_soon(too_slow)
        result = await trio_parallel.run_sync(hard_work, 2, None)
        nursery.cancel_scope.cancel()
    print("got", result, "in", time.perf_counter() - t0, "seconds")


if __name__ == "__main__":
    multiprocessing.freeze_support()
    trio.run(amain)

Documentation

The full API is documented at https://trio-parallel.readthedocs.io/

Features

  • Bypasses the GIL for CPU bound work

  • Minimal API complexity (looks and feels like Trio threads)

  • Cross-platform

  • Automatic LIFO caching of subprocesses

  • Cancel seriously misbehaving code

    • currently via SIGKILL/TerminateProcess

  • Convert segfaults and other scary things to catchable errors

This project aims to use the lightest-weight, lowest-overhead, lowest latency method to achieve CPU parallelism of arbitrary Python code. At the moment, that means subprocesses. However, this project is not at all constrained by that, and will be considering subinterpreters, or any other avenue as they become available.

Currently, this project is based on multiprocessing has all the usual multiprocessing caveats (freeze_support, pickleable objects only). The case for basing these workers on multiprocessing is that it keeps a lot of complexity outside of the project while offering a set of quirks that users are likely already familiar with.

FAQ

Can I have my workers talk to each other?

This is currently possible through the use of multiprocessing.Manager, but we don’t and will not support it. Instead, try using trio.run_process and having the various Trio runs talk to each other over sockets. Also, look into tractor?

Can I let my workers outlive the main Trio process?

The worker processes are started with the daemon flag for lifetime management, so this use case is not supported.

How should I map a function over a collection of arguments?

This is fully possible but we leave the implementation of that up to you. Also, look into trimeter?

Contributing

If you notice any bugs, need any help, or want to contribute any code, GitHub issues and pull requests are very welcome! Please read the code of conduct.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

trio-parallel-0.2.0.tar.gz (26.3 kB view hashes)

Uploaded Source

Built Distribution

trio_parallel-0.2.0-py3-none-any.whl (20.8 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page