[CI] enable custom ops build (#466)

### What this PR does / why we need it?
This PR enable custom ops build  by default. 

### Does this PR introduce _any_ user-facing change?

Yes, users now install vllm-ascend from source will trigger custom ops
build step.

### How was this patch tested?
By image build and e2e CI

---------

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

tests/multicard/test_offline_inference_distributed.py

@@ -0,0 +1,55 @@
+#
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+# This file is a part of the vllm-ascend project.
+# Adapted from vllm/tests/basic_correctness/test_basic_correctness.py
+# Copyright 2023 The vLLM team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+"""Compare the short outputs of HF and vLLM when using greedy sampling.
+
+Run `pytest tests/test_offline_inference.py`.
+"""
+import os
+
+import pytest
+import vllm  # noqa: F401
+from conftest import VllmRunner
+
+os.environ["PYTORCH_NPU_ALLOC_CONF"] = "max_split_size_mb:256"
+
+
+@pytest.mark.parametrize("model, distributed_executor_backend", [
+    ("Qwen/QwQ-32B", "mp"),
+])
+def test_models_distributed(model: str,
+                            distributed_executor_backend: str) -> None:
+    example_prompts = [
+        "vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs.",
+        "Briefly describe the major milestones in the development of artificial intelligence from 1950 to 2020.",
+        "Compare and contrast artificial intelligence with human intelligence in terms of processing information.",
+    ]
+    dtype = "half"
+    max_tokens = 5
+    with VllmRunner(
+            model,
+            dtype=dtype,
+            tensor_parallel_size=4,
+            distributed_executor_backend=distributed_executor_backend,
+    ) as vllm_model:
+        vllm_model.generate_greedy(example_prompts, max_tokens)
+
+
+if __name__ == "__main__":
+    import pytest
+    pytest.main([__file__])

tests/singlecard/test_offline_inference.py

@@ -53,28 +53,6 @@ def test_models(model: str, dtype: str, max_tokens: int) -> None:
         vllm_model.generate_greedy(example_prompts, max_tokens)
 
 
-@pytest.mark.multinpu
-@pytest.mark.parametrize("model, distributed_executor_backend", [
-    ("Qwen/QwQ-32B", "mp"),
-])
-def test_models_distributed(vllm_runner, model: str,
-                            distributed_executor_backend: str) -> None:
-    example_prompts = [
-        "vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs.",
-        "Briefly describe the major milestones in the development of artificial intelligence from 1950 to 2020.",
-        "Compare and contrast artificial intelligence with human intelligence in terms of processing information.",
-    ]
-    dtype = "half"
-    max_tokens = 5
-    with vllm_runner(
-            model,
-            dtype=dtype,
-            tensor_parallel_size=4,
-            distributed_executor_backend=distributed_executor_backend,
-    ) as vllm_model:
-        vllm_model.generate_greedy(example_prompts, max_tokens)
-
-
 if __name__ == "__main__":
     import pytest
     pytest.main([__file__])