[torch.compile] add a flag to track batchsize statistics (#11059)

Signed-off-by: youkaichao <youkaichao@gmail.com>

vllm/envs.py

@@ -69,6 +69,7 @@ if TYPE_CHECKING:
     VLLM_DISABLED_KERNELS: List[str] = []
     VLLM_USE_V1: bool = False
     VLLM_ENABLE_V1_MULTIPROCESSING: bool = False
+    VLLM_LOG_BATCHSIZE_INTERVAL: float = -1
 
 
 def get_default_cache_root():
@@ -452,6 +453,8 @@ environment_variables: Dict[str, Callable[[], Any]] = {
     # If set, enable multiprocessing in LLM for the V1 code path.
     "VLLM_ENABLE_V1_MULTIPROCESSING":
     lambda: bool(int(os.getenv("VLLM_ENABLE_V1_MULTIPROCESSING", "0"))),
+    "VLLM_LOG_BATCHSIZE_INTERVAL":
+    lambda: float(os.getenv("VLLM_LOG_BATCHSIZE_INTERVAL", "-1")),
 }
 
 # end-env-vars-definition