[MISC] Consolidate cleanup() and refactor offline_inference_with_prefix.py (#9510)

tests/tensorizer_loader/conftest.py

@@ -1,27 +1,18 @@
-import contextlib
 import functools
 import gc
 from typing import Callable, TypeVar
 
 import pytest
-import ray
 import torch
 from typing_extensions import ParamSpec
 
-from vllm.distributed import (destroy_distributed_environment,
-                              destroy_model_parallel)
+from vllm.distributed import cleanup_dist_env_and_memory
 from vllm.model_executor.model_loader.tensorizer import TensorizerConfig
 
 
 @pytest.fixture(autouse=True)
 def cleanup():
-    destroy_model_parallel()
-    destroy_distributed_environment()
-    with contextlib.suppress(AssertionError):
-        torch.distributed.destroy_process_group()
-    ray.shutdown()
-    gc.collect()
-    torch.cuda.empty_cache()
+    cleanup_dist_env_and_memory(shutdown_ray=True)
 
 
 _P = ParamSpec("_P")