Skip to content

Add unique prefix - increasing counter #217

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

MML-coder
Copy link
Collaborator

The SyntheticTextItemsGenerator was generating prompts that could trigger vLLM's automatic prefix caching, leading to hitting the prefix cache up to 80% in some cases during the performance benchmarking.

Implemented unique prefix injection to guarantee 0% prefix cache hit rate while maintaining realistic prompt characteristics.

Test:
Performing some tests on the H200 target accelerator to confirm the fix.

📦 Build Artifacts Available
The build artifacts (.whl and .tar.gz) have been successfully generated and are available for download: https://github.com/neuralmagic/guidellm/actions/runs/16009951836/artifacts/3444362586.
They will be retained for up to 30 days.

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.

1 participant