[perf]Qwen3-Omni performance optimization (#3203)

Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>

tests/dfx/perf/scripts/run_benchmark.py

@@ -65,11 +65,13 @@ def omni_server(request):
     Multi-stage initialization can take 10-20+ minutes.
     """
     with _omni_server_lock:
-        test_name, model, stage_config_path, stage_overrides, extra_cli_args = request.param
+        test_name, model, stage_config_path, stage_overrides, extra_cli_args, use_omni = request.param
 
         print(f"Starting OmniServer with test: {test_name}, model: {model}")
 
-        server_args = ["--stage-init-timeout", "600", "--init-timeout", "900"]
+        server_args: list[str] = []
+        if use_omni:
+            server_args += ["--stage-init-timeout", "600", "--init-timeout", "900"]
         # --deploy-config and --stage-overrides compose at the CLI (see vllm_omni/entrypoints/utils.py):
         # deploy-config sets the base; stage-overrides are applied on top. Both can be set.
         if stage_config_path:
@@ -78,7 +80,7 @@ def omni_server(request):
             server_args = ["--stage-overrides", stage_overrides] + server_args
         if extra_cli_args:
             server_args = list(extra_cli_args) + server_args
-        with OmniServer(model, server_args) as server:
+        with OmniServer(model, server_args, use_omni=use_omni) as server:
             server.test_name = test_name
             print("OmniServer started successfully")
             yield server