[Perf] Optimize _update_states for GPU model runner (vllm-project#16910)

Signed-off-by: snowcharm <[email protected]>
Signed-off-by: Agata Dobrzyniewicz <[email protected]>

Author: SnowCharmQ

vllm/v1/worker/gpu_model_runner.py

@@ -454,7 +454,7 @@ def _update_states(self, scheduler_output: "SchedulerOutput") -> None:
 
         # Add the new or resumed requests to the persistent batch.
         # The smaller empty indices are filled first.
-        removed_req_indices = sorted(removed_req_indices, reverse=True)
+        removed_req_indices.sort(reverse=True)
         for req_id in req_ids_to_add:
             req_state = self.requests[req_id]
             if removed_req_indices: