Revert "drop ascend scheduler" (#4580)

Reverts vllm-project/vllm-ascend#4498
- vLLM version: v0.11.2
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.2

tests/e2e/nightly/models/test_deepseek_r1_0528_w8a8.py

@@ -33,6 +33,7 @@ MODES = [
     "single",
     "aclgraph",
     "aclgraph_mlapo",
+    "no_chunkprefill",
 ]
 
 prompts = [
@@ -81,6 +82,9 @@ async def test_models(model: str, mode: str) -> None:
         "method": "deepseek_mtp"
     }
     additional_config = {
+        "ascend_scheduler_config": {
+            "enabled": False
+        },
         "torchair_graph_config": {
             "enabled": True,
             "enable_multistream_moe": False,
@@ -108,6 +112,10 @@ async def test_models(model: str, mode: str) -> None:
     if mode == "aclgraph_mlapo":
         env_dict["VLLM_ASCEND_ENABLE_MLAPO"] = "1"
         additional_config["torchair_graph_config"] = {"enabled": False}
+    if mode == "no_chunkprefill":
+        additional_config["ascend_scheduler_config"] = {"enabled": True}
+        i = server_args.index("--max-num-batched-tokens") + 1
+        server_args[i] = "36864"
     server_args.extend(["--additional-config", json.dumps(additional_config)])
     request_keyword_args: dict[str, Any] = {
         **api_keyword_args,

tests/e2e/nightly/models/test_deepseek_r1_w8a8_eplb.py

@@ -71,6 +71,9 @@ async def test_models(model: str) -> None:
         "cudagraph_mode": "FULL_DECODE_ONLY"
     }
     additional_config: dict[str, Any] = {
+        "ascend_scheduler_config": {
+            "enabled": False
+        },
         "torchair_graph_config": {
             "enabled": True
         },

tests/e2e/nightly/models/test_deepseek_v3_2_exp_w8a8.py

@@ -92,6 +92,7 @@ async def test_models(model: str, tp_size: int, dp_size: int,
         "--gpu-memory-utilization",
         "0.9",
         "--additional-config",
+        '{"ascend_scheduler_config":{"enabled":true},'
         '"torchair_graph_config":{"enabled":true,"graph_batch_sizes":[16]}}',
     ]
     if full_graph:

tests/e2e/nightly/models/test_qwen2_5_vl_32b.py

@@ -85,8 +85,9 @@ async def test_models(model: str, tp_size: int) -> None:
         str(tp_size), "--port",
         str(port), "--max-model-len", "30000", "--max-num-batched-tokens",
         "40000", "--max-num-seqs", "400", "--trust-remote-code",
-        "--gpu-memory-utilization", "0.8", "--compilation_config",
-        '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
+        "--gpu-memory-utilization", "0.8", "--additional-config",
+        '{"ascend_scheduler_config":{"enabled":false}}',
+        "--compilation_config", '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
     ]
     request_keyword_args: dict[str, Any] = {
         **api_keyword_args,

tests/e2e/nightly/models/test_qwen3_235b_a22b_w8a8_eplb.py

@@ -60,7 +60,11 @@ async def test_models(model: str) -> None:
         "PYTORCH_NPU_ALLOC_CONF": "expandable_segments:True",
         "VLLM_ASCEND_ENABLE_FLASHCOMM1": "1"
     }
-    additional_config: dict[str, Any] = {}
+    additional_config: dict[str, Any] = {
+        "ascend_scheduler_config": {
+            "enabled": False
+        },
+    }
     compilation_config = {"cudagraph_mode": "FULL_DECODE_ONLY"}
     server_args = [
         "--quantization", "ascend", "--async-scheduling",

tests/e2e/nightly/models/test_qwen3_235b_w8a8.py

@@ -63,6 +63,11 @@ async def test_models(model: str, mode: str) -> None:
         "PYTORCH_NPU_ALLOC_CONF": "expandable_segments:True",
         "VLLM_ASCEND_ENABLE_FLASHCOMM1": "1"
     }
+    additional_config: dict[str, Any] = {
+        "ascend_scheduler_config": {
+            "enabled": False
+        },
+    }
     compilation_config = {"cudagraph_mode": "FULL_DECODE_ONLY"}
     server_args = [
         "--quantization", "ascend", "--async-scheduling",
@@ -77,6 +82,7 @@ async def test_models(model: str, mode: str) -> None:
     server_args.extend(
         ["--compilation-config",
          json.dumps(compilation_config)])
+    server_args.extend(["--additional-config", json.dumps(additional_config)])
     request_keyword_args: dict[str, Any] = {
         **api_keyword_args,
     }

tests/e2e/nightly/models/test_qwq_32b.py

@@ -93,6 +93,8 @@ async def test_models(model: str, mode: str, tp_size: int) -> None:
         server_args.remove(
             '{"cudagraph_mode":"FULL_DECODE_ONLY", "cudagraph_capture_sizes": [1, 8, 24, 48, 60]}'
         )
+        server_args.append("--additional-config")
+        server_args.append('{"ascend_scheduler_config":{"enabled":true}}')
         server_args.append("--enforce-eager")
     request_keyword_args: dict[str, Any] = {
         **api_keyword_args,