don't enable AscendScheduler when async_scheduling is enabled

Signed-off-by: Ronald1995 <ronaldautomobile@163.com>

Author: Ronald1995

vllm_ascend/platform.py

@@ -135,7 +135,8 @@ def check_and_update_config(cls, vllm_config: VllmConfig) -> None:
         else:
             structured_outputs_config = vllm_config.structured_outputs_config
 
-        if model_config is not None and not model_config.use_mla:
+        if (model_config is not None and not model_config.use_mla
+                and not scheduler_config.async_scheduling):
             logger.info(
                 "Non-MLA LLMs forcibly disable the chunked prefill feature,"
                 "as the performance of operators supporting this feature "