From 0f71e24034263363e48cdc6ae036e7ca057a4f44 Mon Sep 17 00:00:00 2001 From: Lukas Geiger Date: Sat, 31 May 2025 02:30:15 +0100 Subject: [PATCH] [Docs] Correct multiprocessing design doc (#18964) Signed-off-by: Lukas Geiger --- docs/design/multiprocessing.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/docs/design/multiprocessing.md b/docs/design/multiprocessing.md index 412c42fd58..4d58fae20f 100644 --- a/docs/design/multiprocessing.md +++ b/docs/design/multiprocessing.md @@ -22,13 +22,13 @@ This document describes how vLLM deals with these challenges. [Python multiprocessing methods](https://docs.python.org/3/library/multiprocessing.html#contexts-and-start-methods) include: -- `spawn` - spawn a new Python process. This will be the default as of Python - 3.14. In macOS, this is already the default. +- `spawn` - spawn a new Python process. The default on Windows and macOS. -- `fork` - Use `os.fork()` to fork the Python interpreter. This is the default - in Python versions prior to 3.14. +- `fork` - Use `os.fork()` to fork the Python interpreter. The default on + Linux for Python versions prior to 3.14. - `forkserver` - Spawn a server process that will fork a new process on request. + The default on Linux for Python version 3.14 and newer. ### Tradeoffs