drop ascend scheduler (#4498)

Ascend scheduler was added for non chunk prefill case before, since that
the npu ops didn't work well with chunked prefill.

Now the ops with chunked prefill work better, it's time to remove the
ascend scheduler to use vLLM default scheduler.

- vLLM version: v0.11.2

---------

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>

tests/ut/test_ascend_config.py

@@ -56,9 +56,6 @@ class TestAscendConfig(TestBase):
         self.assertTrue(torchair_graph_config.enable_frozen_parameter)
         self.assertFalse(torchair_graph_config.enable_kv_nz)
 
-        ascend_scheduler_config = ascend_config.ascend_scheduler_config
-        self.assertFalse(ascend_scheduler_config.enabled)
-
     @_clean_up_ascend_config
     def test_init_ascend_config_with_additional_config(self):
         test_vllm_config = VllmConfig()
@@ -74,9 +71,6 @@ class TestAscendConfig(TestBase):
                 "enable_kv_nz": True
             },
             "multistream_overlap_shared_expert": True,
-            "ascend_scheduler_config": {
-                "enabled": True
-            },
             "expert_map_path": "test_expert_map_path",
             "refresh": True,
         }
@@ -94,9 +88,6 @@ class TestAscendConfig(TestBase):
         self.assertTrue(torchair_graph_config.enable_frozen_parameter)
         self.assertTrue(torchair_graph_config.enable_kv_nz)
 
-        ascend_scheduler_config = ascend_config.ascend_scheduler_config
-        self.assertTrue(ascend_scheduler_config.enabled)
-
     @_clean_up_ascend_config
     def test_init_ascend_config_with_refresh(self):
         test_vllm_config = VllmConfig()