[CI] Add DeepSeek R1 W8A8 HMB nightly ci (#5874)

### What this PR does / why we need it?

Add DeepSeek R1 W8A8 HMB nightly ci

- vLLM version: v0.13.0
- vLLM main:
bde38c11df

Signed-off-by: hfadzxy <starmoon_zhang@163.com>

.github/workflows/nightly_test_a3.yaml

@@ -154,6 +154,9 @@ jobs:
           # - name: deepseek3_2-exp-w8a8
           #   os: linux-aarch64-a3-16
           #   tests: tests/e2e/nightly/single_node/models/test_deepseek_v3_2_exp_w8a8.py
+          - name: deepseek-r1-w8a8-hmb
+            os: linux-aarch64-a3-16
+            tests: tests/e2e/nightly/single_node/models/test_deepseek_r1_w8a8_hmb.py
     uses: ./.github/workflows/_e2e_nightly_single_node.yaml
     with:
       vllm: v0.13.0

tests/e2e/nightly/single_node/models/test_deepseek_r1_w8a8_hmb.py

@@ -0,0 +1,123 @@
+# 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.
+# This file is a part of the vllm-ascend project.
+#
+import json
+from typing import Any
+
+import openai
+import pytest
+from vllm.utils.network_utils import get_open_port
+
+from tests.e2e.conftest import RemoteOpenAIServer
+from tools.aisbench import run_aisbench_cases
+
+MODELS = [
+    "vllm-ascend/DeepSeek-R1-W8A8",
+]
+
+MODES = [
+    "single",
+]
+
+prompts = [
+    "San Francisco is a",
+]
+
+api_keyword_args = {
+    "max_tokens": 10,
+}
+
+aisbench_cases = [{
+    "case_type": "accuracy",
+    "dataset_path": "vllm-ascend/gsm8k-lite",
+    "request_conf": "vllm_api_general_chat",
+    "dataset_conf": "gsm8k/gsm8k_gen_0_shot_cot_chat_prompt",
+    "max_out_len": 6000,
+    "batch_size": 32,
+    "baseline": 95,
+    "threshold": 5
+}, {
+    "case_type": "performance",
+    "dataset_path": "vllm-ascend/GSM8K-in3500-bs400",
+    "request_conf": "vllm_api_stream_chat",
+    "dataset_conf": "gsm8k/gsm8k_gen_0_shot_cot_str_perf",
+    "num_prompts": 32,
+    "max_out_len": 1500,
+    "batch_size": 32,
+    "baseline": 1,
+    "threshold": 0.97
+}]
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize("model", MODELS)
+@pytest.mark.parametrize("mode", MODES)
+async def test_models(model: str, mode: str) -> None:
+    port = get_open_port()
+    env_dict = {
+        "HCCL_BUFFSIZE": "1024",
+    }
+
+    additional_config = {
+        "ascend_scheduler_config": {
+            "enabled": False
+        },
+        "torchair_graph_config": {
+            "enabled": False,
+            "enable_multistream_shared_expert": False
+        }
+    }
+
+    server_args = [
+        "--quantization", "ascend", "--port",
+        str(port), "--data-parallel-size", "8", "--data-parallel-size-local",
+        "8", "--data-parallel-rpc-port", "13389", "--tensor-parallel-size",
+        "2", "--enable-expert-parallel", "--seed", "1024", "--max-num-seqs",
+        "32", "--max-model-len", "6000", "--max-num-batched-tokens", "6000",
+        "--trust-remote-code", "--gpu-memory-utilization", "0.92",
+        "--no-enable-prefix-caching", "--reasoning-parser", "deepseek_r1"
+    ]
+
+    if mode == "single":
+        server_args.append("--enforce-eager")
+
+    server_args.extend(["--additional-config", json.dumps(additional_config)])
+
+    request_keyword_args: dict[str, Any] = {
+        **api_keyword_args,
+    }
+
+    with RemoteOpenAIServer(model,
+                            server_args,
+                            server_port=port,
+                            env_dict=env_dict,
+                            auto_port=False) as server:
+        client = server.get_async_client()
+        batch = await client.completions.create(
+            model=model,
+            prompt=prompts,
+            **request_keyword_args,
+        )
+        choices: list[openai.types.CompletionChoice] = batch.choices
+        assert choices[0].text, "empty response"
+        print(choices)
+        # aisbench test
+        if mode in ["single"]:
+            return
+        run_aisbench_cases(model,
+                           port,
+                           aisbench_cases,
+                           server_args=server_args)