Add topology-aware qwen27b 0-8k tuning

configs/examples/dash0_qwen27b_tight_slo_run4_0_8k.json

@@ -46,6 +46,7 @@
       "VLLM_ATTENTION_BACKEND": "FLASH_ATTN",
       "VLLM_QUANTIZE_ROUTED_EXPERTS_ONLY": "1",
       "VLLM_USE_FLASHINFER_SAMPLER": "0",
+      "VLLM_DP_MASTER_PORT": "9528",
       "VLLM_RESPONSE_TIMEOUT": "300",
       "VLLM_LOG_REQ_KV_LENS": "1",
       "DS_LLM_GRACEFUL_SHUTDOWN_KEEP_SECONDS": "600"
@@ -73,25 +74,36 @@
       "skip-mm-profiling": true,
       "quantization": "fp8",
       "tensor-parallel-size": 4,
+      "data-parallel-size": 1,
+      "expert-parallel-size": 1,
       "max-num-seqs": 16,
       "disable-log-requests": true
     },
     "tunable_envs": [
-      "VLLM_ATTENTION_BACKEND",
-      "VLLM_ENABLE_TORCH_COMPILE",
-      "VLLM_USE_FLASHINFER_SAMPLER",
-      "VLLM_ENABLE_MODEL_RUNNER_WARMUP"
+      "VLLM_ENABLE_TORCH_COMPILE"
     ],
     "tunable_flags": [
       "tensor-parallel-size",
+      "data-parallel-size",
+      "expert-parallel-size",
       "gpu-memory-utilization",
       "block-size",
       "max-num-batched-tokens",
       "max-num-seqs",
       "enable-prefix-caching",
-      "enable-chunked-prefill",
-      "disable-cascade-attn"
+      "enable-chunked-prefill"
     ],
+    "topology_constraints": {
+      "require_tp_dp_product_equals_gpu_count": false,
+      "require_ep_size_leq_tp_dp_product": true,
+      "require_ep_size_divides_tp_dp_product": true,
+      "require_enable_expert_parallel_when_ep_gt_one": true,
+      "validate_cuda_graph_sizes_divisible_by_tp_when_tp_ep_reduce_scatter": true,
+      "allowed_tp_dp_products": [1, 2, 4, 8],
+      "allowed_tensor_parallel_sizes": [1, 2, 4],
+      "allowed_data_parallel_sizes": [1, 2, 4, 8],
+      "allowed_expert_parallel_sizes": [1]
+    },
     "python_executable": "python3"
   },
   "trace": {
@@ -144,6 +156,7 @@
     "endpoint": {
       "provider": "codex",
       "model": "gpt-5.4",
+      "stream": true,
       "api_key_env": "OPENAI_API_KEY",
       "timeout_s": 180
     }