EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

[CI] tune the test order to warmup the server (#5860)

Browse Source
This commit is contained in:
Lianmin Zheng
2025-04-28 19:27:37 -07:00
committed by GitHub
parent 1cc326032d
commit 26fc32d168
2 changed files with 7 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
python/sglang/srt/layers/quantization/deep_gemm.py
View File

@@ -293,7 +293,7 @@ def _maybe_compile_deep_gemm_one_type_all(
logger.info( logger.info(
f"Try DeepGEMM JIT Compiling for " f"Try DeepGEMM JIT Compiling for "
f"<{kernel_helper.name}> N={n}, K={k}, num_groups={num_groups} with all Ms." f"<{kernel_helper.name}> N={n}, K={k}, num_groups={num_groups} with all Ms."
f"{' It only takes a litte time(Typically 1 sec) if you have run `sglang.compile_deep_gemm`. ' if not _IN_PRECOMPILE_STAGE else ''}" f"{' It only takes a litte time (typically 1 sec) if you have run `python3 -m sglang.compile_deep_gemm`. ' if not _IN_PRECOMPILE_STAGE else ''}"
) )
# NOTE(alcanderian): get_num_sms should be change when 2-batch-overlap is introduced # NOTE(alcanderian): get_num_sms should be change when 2-batch-overlap is introduced

8
test/srt/test_full_deepseek_v3.py
View File

@@ -35,7 +35,9 @@ class TestDeepseekV3(CustomTestCase):
def tearDownClass(cls): def tearDownClass(cls):
kill_process_tree(cls.process.pid) kill_process_tree(cls.process.pid)
def test_gsm8k(self): def test_a_gsm8k(
self,
): # Append an "a" to make this test run first (alphabetically) to warm up the server
args = SimpleNamespace( args = SimpleNamespace(
num_shots=8, num_shots=8,
data_path=None, data_path=None,
@@ -100,7 +102,9 @@ class TestDeepseekV3MTP(CustomTestCase):
def tearDownClass(cls): def tearDownClass(cls):
kill_process_tree(cls.process.pid) kill_process_tree(cls.process.pid)
def test_gsm8k(self): def test_a_gsm8k(
self,
): # Append an "a" to make this test run first (alphabetically) to warm up the server
requests.get(self.base_url + "/flush_cache") requests.get(self.base_url + "/flush_cache")
args = SimpleNamespace( args = SimpleNamespace(