# ==========================================
# ACTUAL TEST CASES
# ==========================================

test_cases:
  - name: "Qwen3-32B-TP4"
    model: "Qwen/Qwen3-32B"
    envs:
      TASK_QUEUE_ENABLE: "1"
      OMP_PROC_BIND: "false"
      HCCL_OP_EXPANSION_MODE: "AIV"
      PAGED_ATTENTION_MASK_LEN: "5500"
      SERVER_PORT: "DEFAULT_PORT"
    server_cmd:
      - "--no-enable-prefix-caching"
      - "--tensor-parallel-size"
      - "4"
      - "--port"
      - "$SERVER_PORT"
      - "--max-model-len"
      - "36864"
      - "--max-num-batched-tokens"
      - "36864"
      - "--block-size"
      - "128"
      - "--trust-remote-code"
      - "--gpu-memory-utilization"
      - "0.9"
      - "--additional-config"
      - '{"enable_weight_nz_layout":true}'
    benchmarks:
      acc:
        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: 32768
        batch_size: 32
        baseline: 95
        threshold: 5
      perf:
        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: 80
        max_out_len: 1500
        batch_size: 20
        request_rate: 0
        baseline: 1
        threshold: 0.97