[shen-shanshan]

No description provided.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[comaniac]

I checked the example again and listed all typos. Probably because I updated the figures but didn't update the text when polishing the design doc. Please help fix all of them in this PR. Thanks!

Time 1: The cache is empty and a new request comes in. We allocate 4 blocks. 3 of them are already full and cached. The fourth block is partially full with 3 of 4 tokens.

Time 4: Request 1 comes in with the 14 prompt tokens, where the first 10 tokens are the same as request 0. We can see that only the first 2 blocks (8 tokens) hit the cache, because the 3rd block only matches 2 of 4 tokens.

[shen-shanshan]

I checked the example again and listed all typos. Probably because I updated the figures but didn't update the text when polishing the design doc. Please help fix all of them in this PR. Thanks!

Time 1: The cache is empty and a new request comes in. We allocate 4 blocks. 3 of them are already full and cached. The fourth block is partially full with 3 of 4 tokens.

Time 4: Request 1 comes in with the 14 prompt tokens, where the first 10 tokens are the same as request 0. We can see that only the first 2 blocks (8 tokens) hit the cache, because the 3rd block only matches 2 of 4 tokens.

OK, I will modify it soon.

[comaniac]

LGTM