[TRTLLM-8366][feat] add kimi multi nodes case (NVIDIA#8025)

Signed-off-by: Xin He (SW-GPU) <[email protected]>

Author: xinhe-nv

tests/integration/defs/test_e2e.py

@@ -3406,7 +3406,7 @@ def test_ptp_quickstart_advanced_llama_multi_nodes(llm_root, llm_venv,
     check_call(" ".join(run_cmd), shell=True, env=llm_venv._new_env)
 
 
-@pytest.mark.timeout(5400)
+@pytest.mark.timeout(7200)
 @pytest.mark.skip_less_device_memory(80000)
 @pytest.mark.skip_less_device(4)
 @pytest.mark.parametrize("eval_task", ["mmlu"])
@@ -3423,6 +3423,8 @@ def test_ptp_quickstart_advanced_llama_multi_nodes(llm_root, llm_venv,
     pytest.param('Qwen3/saved_models_Qwen3-235B-A22B_nvfp4_hf',
                  marks=skip_pre_blackwell),
     pytest.param('DeepSeek-R1/DeepSeek-R1-0528-FP4', marks=skip_pre_blackwell),
+    pytest.param('Kimi-K2-Instruct',
+                 marks=(skip_pre_hopper, skip_post_blackwell)),
 ])
 def test_multi_nodes_eval(llm_venv, model_path, tp_size, pp_size, ep_size,
                           eval_task):

tests/integration/test_lists/qa/llm_function_multinode.txt

@@ -7,5 +7,6 @@ test_e2e.py::test_multi_nodes_eval[llama4-models/nvidia/Llama-4-Maverick-17B-128
 test_e2e.py::test_multi_nodes_eval[Qwen3/Qwen3-235B-A22B-tp16-mmlu]
 test_e2e.py::test_multi_nodes_eval[Qwen3/saved_models_Qwen3-235B-A22B_nvfp4_hf-tp16-mmlu]
 test_e2e.py::test_multi_nodes_eval[DeepSeek-R1/DeepSeek-R1-0528-FP4-tp16-mmlu]
+test_e2e.py::test_multi_nodes_eval[Kimi-K2-Instruct-tp16-mmlu]
 test_e2e.py::test_openai_disagg_multi_nodes_completion[ctx_tp2pp1-gen_tp2pp1]
 test_e2e.py::test_openai_disagg_multi_nodes_completion[ctx_tp1pp2-gen_tp1pp2]