Skip to content

Conversation

@happyandslow
Copy link
Collaborator

Pull Request Description

This PR fixes multiple existing issues with benchmark implementation

Related Issues

Resolves: #1029

@happyandslow happyandslow marked this pull request as draft May 5, 2025 20:56
@happyandslow happyandslow marked this pull request as ready for review May 7, 2025 17:41
--model \"$TARGET_MODEL\" \
--api-key \"$API_KEY\" \
--time-scale \"$TIME_SCALE\" \
--routing-strategy \"$ROUTING_STRATEGY\" \
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

if we plan to use benchmark against other framework, this is not needed.

@Jeffwan
Copy link
Collaborator

Jeffwan commented May 7, 2025

Overall looks good to me. @happyandslow can you or du help on this issue so we can test these code in each CI run. #1050

@Jeffwan Jeffwan merged commit 2b95773 into vllm-project:main May 8, 2025
3 checks passed
Yaegaki1Erika pushed a commit to Yaegaki1Erika/aibrix that referenced this pull request Jul 23, 2025
…1049)

* move prompt decoration to client dispatch function; remove locks
* update multiturn dataset generation by adding shared prefix length
* update README for workload output format
* move predefined synthetic workload files to autoscaling scenarios
* update figure explaination in workload generator README
* using shuffling instead of sampling from request finder
* Adding stats to workload generation
* adding routing strategy knob
* bug fix in client.py
* print fix in workload generator
---------

Signed-off-by: Le Xu <[email protected]>
Co-authored-by: Le Xu <[email protected]>
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.

Tracking multiple benchmark code issues

2 participants