NVIDIA / NVIDIA/TensorRT-LLM

trtllm-serve --help takes 11 seconds to respond

Open
#3,587 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
14.7k
Forks
2.8k
Avg merge
2d 23h
Merged PRs (30d)
489

Description

This is causing users to think this command timed out.
`user@031d6fd4c05c:/workspace$ time trtllm-serve --help
2025-04-15 20:31:19,154 - INFO - flashinfer.jit: Prebuilt kernels not found, using JIT backend
[TensorRT-LLM] TensorRT-LLM version: 0.19.0.dev2025041500
Usage: trtllm-serve [OPTIONS] COMMAND [ARGS]...

Options:
--help Show this message and exit.

Commands:
disaggregated Running server in disaggregated mode
disaggregated_mpi_worker Launching disaggregated MPI worker
serve Running an OpenAI API compatible server

real 0m11.451s
user 0m18.162s
sys 0m9.453s
`

Need to improve this for usability

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 running the trtllm-serve --help entry point and measuring where its 11-second startup is spent, using the command output in this issue as the baseline. Trace the startup work that occurs before help is displayed, then verify that help responds substantially faster without changing the listed commands or options.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
cli, performance
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.