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
remove duplicate code within cleanup_dist_env_and_memory (#22147)
Signed-off-by: Andy Xie <andy.xning@gmail.com>
Ning Xie committed
Aug 4, 2025 at 11:03 UTC
c2e75b3c11047eec0f184577ce134879ce993f77
1 file changed
-2
vllm/distributed/parallel_state.py
-2
@@ -1238,8 +1238,6 @@ def destroy_distributed_environment():
1238
def cleanup_dist_env_and_memory(shutdown_ray: bool = False):
1239
destroy_model_parallel()
1240
destroy_distributed_environment()
1241
- with contextlib.suppress(AssertionError):
1242
- torch.distributed.destroy_process_group()
1241
if shutdown_ray:
1242
import ray # Lazy import Ray
1243
ray.shutdown()