NVIDIA / NVIDIA/TensorRT-LLM

[Feature]: AutoDeploy: Remove host sync for MTP + overlap scheduler + mixed prefill/decode batches

Open
#15,289 0 comments 0 reactions 1 assignee View on GitHub

Nobody has claimed this yet.

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

There was a host sync in the above setting, caused by the fact the overlap scheduler was trying to make adjustments to the input_pos of prefill sequences of unknown length. This can be fixed by simply noticing that prefill sequences do not have prior-iteration MTP acceptance to factor in, so they don't need any adjustment in the overlap scheduler and can be skipped.

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.