A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 01:00 UTC
[Bugfix] Fix eviction cached blocked logic (#21357)
Signed-off-by: simon-mo <simon.mo@hey.com>
Simon Mo committed
Jul 22, 2025 at 01:18 UTC
32142b3c62277ac7cb941f2036270decb6b514f4
1 file changed
+1
-1
vllm/v1/core/block_pool.py
+1
-1
@@ -253,7 +253,7 @@ class BlockPool:
253
return False
254
block.reset_hash()
255
blocks_by_id.pop(block.block_id, None)
256
- if blocks_by_id:
256
+ if len(blocks_by_id) == 0:
257
del self.cached_block_hash_to_block[block_hash]
258
259
if self.enable_kv_cache_events: