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

[bugfix][DCP] fix block_size of hash in DCP prefix caching (#26296)

Browse Source 
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
This commit is contained in:
Chen Zhang
2025-10-10 18:02:49 +08:00
committed by GitHub
parent 720d3cd0f0
commit 606b00e80f
5 changed files with 12 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
vllm/v1/engine/core.py
View File

@@ -142,12 +142,18 @@ class EngineCore:
logger.info("Disabling chunked prefill for model without KVCache")
vllm_config.scheduler_config.chunked_prefill_enabled = False
scheduler_block_size = (
vllm_config.cache_config.block_size
* vllm_config.parallel_config.decode_context_parallel_size
)
self.scheduler: SchedulerInterface = Scheduler(
vllm_config=vllm_config,
kv_cache_config=kv_cache_config,
structured_output_manager=self.structured_output_manager,
include_finished_set=vllm_config.parallel_config.data_parallel_size > 1,
log_stats=self.log_stats,
block_size=scheduler_block_size,
)
self.use_spec_decode = vllm_config.speculative_config is not None
if self.scheduler.connector is not None: # type: ignore
@@ -177,14 +183,13 @@ class EngineCore:
self.vllm_config.cache_config.enable_prefix_caching
or self.scheduler.get_kv_connector() is not None
):
block_size = vllm_config.cache_config.block_size
caching_hash_fn = get_hash_fn_by_name(
vllm_config.cache_config.prefix_caching_hash_algo
)
init_none_hash(caching_hash_fn)
self.request_block_hasher = get_request_block_hasher(
block_size, caching_hash_fn
scheduler_block_size, caching_hash_fn
)
self.step_fn = (