sgl-project/sglang

Reducing the time to import sglang

Open

#10,492 opened on Sep 16, 2025

View on GitHub
 (6 comments) (4 reactions) (1 assignee)Python (6,216 forks)auto 404
good first issuehigh priority

Repository metrics

Stars
 (28,442 stars)
PR merge metrics
 (Avg merge 2d 1h) (1,000 merged PRs in 30d)

Description

Now importing sglang takes too long. On my H100 machine, it took 12s to just import sglang. Can someone help reducing the importing time?

> time python -c "from sglang.srt.managers.scheduler import Scheduler"
All deep_gemm operations loaded successfully!
W0916 03:22:46.348000 152559 torch/utils/cpp_extension.py:2425] TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation.
W0916 03:22:46.348000 152559 torch/utils/cpp_extension.py:2425] If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'] to specific architectures.

real	0m12.310s
user	0m13.728s
sys	0m23.912s

Contributor guide

Research direction
Profile the import process to identify which modules take the most time. Consider using lazy imports or delaying imports of heavy dependencies.
Tech stack
pythonpytorch
Domain
backendmachine learningai
Issue type
Performance
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
2
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
Python
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
75