How to run async task inside subinterpreter?

[baseplate-admin]

Hi,
I have this code

import anyio



async def hello_world():
    print("Hello, world!")
    await anyio.sleep(1)
    print("Goodbye, world!")


async def main():
    anyio.to_interpreter.run_sync(hello_world)


anyio.run(main)

But this is not working as expected, is async under subinterpreter under works? Or what i asked here is not possible.

Thanks in advance

[agronholm]

Why would you want to do something like this? The whole point of subinterpreters is to allow taking advantage of multiple CPU cores while running CPU intensive synchronous Python code.

[davidbrochart]

Maybe @baseplate-admin is looking for something like a blocking portal that would make use of a subinterpreter instead of a thread?

[agronholm]

Subinterpreters are run in threads, just so we're clear.

[baseplate-admin]

Why would you want to do something like this? The whole point of subinterpreters is to allow taking advantage of multiple CPU cores while running CPU intensive synchronous Python code.

Basically i have a udp server and a http server I want to run each server in separate subinterpreter. Both are async and the workers doesn't share any data in between them, So i was thinking of spawning two subinterpreters.

This is my current approach for running the app. This is not using all the cores in the system (as expected), so i am thinking maybe subinterpreter solves this issue? Though i am not sure.

Thanks a bunch.

[agronholm]

Perhaps you'd like to use InterpreterPoolExecutor to run separate event loops then? I don't think the AnyIO facilities are the right solution for you. Also be aware of the WIP nature of the subinterpreters interface in Python.

[baseplate-admin]

Perhaps you'd like to use InterpreterPoolExecutor to run separate event loops then? I don't think the AnyIO facilities are the right solution for you. Also be aware of the WIP nature of the subinterpreters interface in Python.

Alright, thanks a bunch for your response.