Support piecewise cuda graph for MLA (#11812)

test/srt/run_suite.py

@@ -103,7 +103,7 @@ suites = {
         TestFile("test_original_logprobs.py", 41),
         TestFile("test_page_size.py", 60),
         TestFile("test_penalty.py", 82),
-        TestFile("test_piecewise_cuda_graph.py", 180),
+        TestFile("test_piecewise_cuda_graph.py", 300),
         TestFile("test_priority_scheduling.py", 130),
         TestFile("test_pytorch_sampling_backend.py", 66),
         TestFile("test_radix_attention.py", 105),

test/srt/test_piecewise_cuda_graph.py

@@ -1,9 +1,11 @@
 import unittest
 
 from sglang.srt.utils import kill_process_tree
+from sglang.test.few_shot_gsm8k import run_eval as run_eval_few_shot_gsm8k
 from sglang.test.run_eval import run_eval
 from sglang.test.test_utils import (
     DEFAULT_MODEL_NAME_FOR_TEST,
+    DEFAULT_MODEL_NAME_FOR_TEST_MLA,
     DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
     DEFAULT_URL_FOR_TEST,
     CustomTestCase,
@@ -92,5 +94,43 @@ class TestPiecewiseCudaGraphQwen3MoE(CustomTestCase):
         self.assertGreaterEqual(metrics["score"], 0.90)
 
 
+class TestPiecewiseCudaGraphDeepSeek(CustomTestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.model = DEFAULT_MODEL_NAME_FOR_TEST_MLA
+        cls.base_url = DEFAULT_URL_FOR_TEST
+        cls.process = popen_launch_server(
+            cls.model,
+            cls.base_url,
+            timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+            other_args=[
+                "--enable-piecewise-cuda-graph",
+                "--piecewise-cuda-graph-compiler",
+                "eager",
+                "--piecewise-cuda-graph-max-tokens",
+                "4096",  # should less than max_context_len
+            ],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+
+    def test_gsm8k(self):
+        args = SimpleNamespace(
+            num_shots=5,
+            data_path=None,
+            num_questions=200,
+            max_new_tokens=512,
+            parallel=128,
+            host="http://127.0.0.1",
+            port=int(self.base_url.split(":")[-1]),
+        )
+        metrics = run_eval_few_shot_gsm8k(args)
+        print(metrics)
+
+        self.assertGreater(metrics["accuracy"], 0.62)
+
+
 if __name__ == "__main__":
     unittest.main()