vllm-project / vllm-project/production-stack

feature: using no GIL or Granian to improve router performance

Open
#309 7 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

feature request
Dominant language
Python
Stars
2.6k
Forks
503
Avg merge
4d 17h
Merged PRs (30d)
8

Description

Describe the feature

We can potentially use Python with No GIL or the Granian HTTP server to improve router performance, throughput, and concurrency.

Why do you need this feature?
  1. Python No GIL: Python 3.13 introduces the possibility to run Python without the GIL. Removing the GIL can drastically improve concurrency in CPU-bound or mixed CPU-/I/O-bound workloads. A few references that show the potential performance gains:
  1. Granian
    Granian is a Rust HTTP server for Python applications. By using Granian as a router backend, we can potentially reduce latency and increase throughput.
Additional context

No response

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by measuring the current router under representative CPU-bound, I/O-bound, and heavy-traffic workloads, then compare Python 3.13 No-GIL and Granian as separate options. Done means the project has reproducible benchmark results and a decided implementation direction, with the affected entry points and validation steps identified.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, rust
Domain
backend, performance
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.