xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 09:16:06 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Log] Warning Once for Cutlass MLA (#23137)

Browse Source 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Wentao Ye 2025-08-19 02:24:16 -04:00 committed by GitHub
parent e75f342261
commit 90bbe0a5ad
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/v1/attention/backends/mla/cutlass_mla.py
View File

@ -115,7 +115,7 @@ class CutlassMLAImpl(MLACommonImpl[MLACommonMetadata]):
self._use_old_cutlass_mla = False
force_old_cutlass = os.environ.get("FORCE_OLD_CUTLASS_MLA", None)
if force_old_cutlass:
logger.warning("Forcing old cutlass mla kernel")
logger.warning_once("Forcing old cutlass mla kernel")
self._use_old_cutlass_mla = True
# TODO: Currently, num_kv_splits is limited to 16 to avoid hanging
@ -123,8 +123,8 @@ class CutlassMLAImpl(MLACommonImpl[MLACommonMetadata]):
# FORCE_NUM_KV_SPLITS=1
force_num_kv_splits = os.environ.get("FORCE_NUM_KV_SPLITS", None)
if force_num_kv_splits:
logger.warning("Forcing num_kv_splits to %d",
int(force_num_kv_splits))
logger.warning_once("Forcing num_kv_splits to %d",
int(force_num_kv_splits))
self._num_kv_splits = int(force_num_kv_splits)
else:
self._num_kv_splits = -1 # => Auto-detect