enable mm allreduce test

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

Author: Ronald1995

tests/e2e/singlecard/test_mm_allreduce.py

@@ -0,0 +1,64 @@
+#
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+# 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.
+#
+
+import os
+import subprocess
+import sys
+from pathlib import Path
+from unittest.mock import patch
+
+import pytest
+import torch_npu
+
+MODELS = ["Qwen/Qwen3-0.6B"]
+DEVICE_NAME = torch_npu.npu.get_device_name(0)[:10]
+
+
+@pytest.mark.skipif(
+    DEVICE_NAME != "Ascend910B",
+    reason="This test is only for Ascend910B devices.",
+)
+@pytest.mark.parametrize("model", MODELS)
+@patch.dict(os.environ, {"VLLM_ASCEND_ENABLE_MATMUL_ALLREDUCE": "1"})
+def test_external_launcher(model):
+    script = Path(
+        __file__
+    ).parent.parent.parent.parent / "examples" / "offline_external_launcher.py"
+    env = os.environ.copy()
+    cmd = [
+        sys.executable,
+        str(script),
+        "--model",
+        model,
+        "--trust-remote-code",
+    ]
+
+    print(f"Running subprocess: {' '.join(cmd)}")
+    proc = subprocess.run(
+        cmd,
+        env=env,
+        stdout=subprocess.PIPE,
+        stderr=subprocess.STDOUT,
+        timeout=600,
+    )
+    output = proc.stdout.decode()
+
+    print(output)
+
+    assert "AscendRowParallelLinear: Matmul all-reduce is enabled" in output
+    assert "Generated text:" in output
+    assert proc.returncode == 0

vllm_ascend/patch/worker/patch_common/patch_linear.py

@@ -25,6 +25,7 @@
 from vllm.distributed import (get_tensor_model_parallel_rank,
                               split_tensor_along_last_dim)
 from vllm.distributed.parallel_state import get_tp_group
+from vllm.logger import logger
 from vllm.model_executor.layers.linear import RowParallelLinear
 
 from vllm_ascend import envs
@@ -142,4 +143,5 @@ def calc_output(self, input_parallel: torch.Tensor) -> torch.Tensor:
 
 
 if envs.VLLM_ASCEND_ENABLE_MATMUL_ALLREDUCE:
+    logger.info("AscendRowParallelLinear: Matmul all-reduce is enabled. ")
     vllm.model_executor.layers.linear.RowParallelLinear = AscendRowParallelLinear