[CI] Tiny enhance the dp attention load blance benchmark (#19194)

python/sglang/test/test_utils.py

@@ -1041,6 +1041,7 @@ def get_benchmark_args(
     gsp_output_len=32,
     gsp_num_turns=1,
     header=None,
+    max_concurrency=None,
 ):
     return SimpleNamespace(
         backend=backend,
@@ -1082,6 +1083,7 @@ def get_benchmark_args(
         gsp_output_len=gsp_output_len,
         gsp_num_turns=gsp_num_turns,
         header=header,
+        max_concurrency=max_concurrency,
     )
 
 

test/registered/distributed/test_disaggregation_dp_attention.py

@@ -1,6 +1,7 @@
 import unittest
 from types import SimpleNamespace
 
+from sglang.bench_serving import run_benchmark
 from sglang.srt.environ import envs
 from sglang.test.ci.ci_register import register_cuda_ci
 from sglang.test.few_shot_gsm8k import run_eval as run_eval_few_shot_gsm8k
@@ -10,11 +11,12 @@ from sglang.test.server_fixtures.disaggregation_fixture import (
 from sglang.test.test_utils import (
     DEFAULT_MODEL_NAME_FOR_TEST_MLA,
     DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+    get_benchmark_args,
     popen_launch_pd_server,
     try_cached_model,
 )
 
-register_cuda_ci(est_time=155, suite="stage-c-test-8-gpu-h20")
+register_cuda_ci(est_time=580, suite="stage-c-test-8-gpu-h20")
 
 
 class TestDisaggregationDPAttention(PDDisaggregationServerBase):
@@ -104,6 +106,24 @@ class TestDisaggregationDPAttention(PDDisaggregationServerBase):
 
 class TestDisaggregationDPAttentionRoundRobin(TestDisaggregationDPAttention):
     LOAD_BALANCE_METHOD = "round_robin"
+    # TODO: add test for other load balance methods
+    # TODO: add a balancedness metric
+
+    def test_bench_serving(self):
+        args = get_benchmark_args(
+            base_url=f"http://{self.base_host}:{self.lb_port}",
+            dataset_name="random",
+            tokenizer=self.model,
+            num_prompts=1000,
+            random_input_len=4096,
+            random_output_len=1024,
+            request_rate=float("inf"),
+            max_concurrency=256,
+        )
+        result = run_benchmark(args)
+
+        self.assertLess(result["mean_tpot_ms"], 20)
+        self.assertEqual(result["completed"], 1000)
 
 
 if __name__ == "__main__":