biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V1][Minor] Remove outdated comment (#12928)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon
2025-02-07 19:07:37 -08:00
committed by GitHub
parent 45cbc4991d
commit b21f0f9d17
1 changed files with 1 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/v1/core/kv_cache_manager.py
View File

@@ -299,9 +299,7 @@ class KVCacheManager:
While all scheduled requests must be in the RUNNING state, the inverse
is not necessarily true. There may be RUNNING requests that are not
scheduled in the current step. As of 1/1/2025, the scheduler does not
allow this case, but it is possible in the future, as we allow more
flexible scheduling.
scheduled in the current step.
This can result in an edge case where the number of common prefix blocks
is 0, even though all scheduled requests share a common prefix. This