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 14, 05:30 UTC
[CI] Fix `tests/distributed/test_ca_buffer_sharing.py` (#22849)
Signed-off-by: ilmarkov <imarkov@redhat.com> Co-authored-by: ilmarkov <imarkov@redhat.com> Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Ilya Markov committed
Aug 14, 2025 at 05:09 UTC
1d20c347179733875faf4c7802ef695ad3dec4f8
1 file changed
+1
-1
vllm/distributed/device_communicators/custom_all_reduce.py
+1
-1
@@ -297,7 +297,7 @@ class CustomAllreduce:
297
@staticmethod
298
def free_shared_buffer(pointers: list[int],
299
group: Optional[ProcessGroup] = None,
300
- rank: Optional[int] = 0) -> None:
300
+ rank: Optional[int] = None) -> None:
301
if rank is None:
302
rank = dist.get_rank(group=group)
303
if ops is not None: