NVIDIA / NVIDIA/TensorRT-LLM

[Feature]: AutoDeploy: investigate latency of first request in trtllm-serve

Open
#9,276 0 comments 0 reactions 1 assignee View on GitHub

@suyoggupta is already working on this.

Since Nov 18, 2025.

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

Description

🚀 The feature, motivation and pitch

from @2ez4bz. Seems that the first request in trtllm-serve or dynamo is much slower than subsequent requests. From logging in dynamo @2ez4bz seems to be able to track it down to intitial calls to flashinfer.

@lucaslie and @2ez4bz discussed whether it may be flashinfer prefill that triggers it since AD doens't do any warmup for prefill

Alternatives

No response

Additional context

No response

Before submitting a new issue...
  • Make sure you already searched for relevant issues, and checked the documentation and examples for answers to frequently asked questions.

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.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.