Skip to content

Conversation

hnyls2002
Copy link
Collaborator

Motivation

Modifications

Checklist

  • Format your code according to the Contributor Guide.
  • Add unit tests as outlined in the Contributor Guide.
  • Update documentation as needed, including docstrings or example tutorials.

@merrymercy merrymercy merged commit ab4a83b into main Sep 5, 2024
9 checks passed
@merrymercy merrymercy deleted the optimize-schedule branch September 5, 2024 21:30
@merrymercy merrymercy mentioned this pull request Sep 13, 2024
29 tasks
@hxer7963
Copy link
Contributor

hi, @hnyls2002 @merrymercy.

I have been exploring the source code of the PrefillAdder class and the scheduler module within ModelTpServer::get_new_prefill_batch. It seems that the implementation reserves the maximum possible output token slots based on the estimated new_token_ratio before scheduling prefill requests.

However, I am curious about the motivation behind the scheduling strategy used by PrefillAdder and how it contributes to optimizing scheduling performance.

Could you provide some insights into these aspects?

timethink pushed a commit to timethink/sglang that referenced this pull request Mar 9, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants