[Runtime] Fix PagedKVCache for PopN and enhance tests #17045
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
tqchen
approved these changes
May 29, 2024
MasterJH5574
force-pushed
the
tvm-dev/2024-05-29-kv-cache-removal-fix
branch
from
d19daf4 to
d052c2d
Compare
This PR fixes a bug in the PagedKVCache which may happen when the sequence removal order is not consistent with the reverse order of sequence add/fork order. With this fix, the PagedKVCache now supports removing sequences in any order without breaking. This PR also adds an `empty` function to PagedKVCache to check if the KV cache is empty. Right now this function is only used for test purpose, where we check if everything in the KV cache is freed after removing all sequences.
MasterJH5574
force-pushed
the
tvm-dev/2024-05-29-kv-cache-removal-fix
branch
from
d052c2d to
46237a1
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR fixes a bug in the PagedKVCache which may happen when the sequence removal order is not consistent with the reverse order of sequence add/fork order. With this fix, the PagedKVCache now supports removing sequences in any order without breaking.
This PR also adds an
emptyfunction to PagedKVCache to check if the KV cache is empty. Right now this function is only used for test purpose, where we check if everything in the KV cache is freed after removing all sequences.