minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

Check the topology identities where the layout is written, and build at the published widths (#40340)

Browse Source
This commit is contained in:
Cheng Wan
2026-09-21 12:22:59 -07:00
committed by GitHub
parent d5fdab7022
commit 2d0e94e3a3
43 changed files with 843 additions and 261 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/manual/ep/test_flashinfer_dispatcher.py
+3 -4
View File
@@ -14,7 +14,7 @@ from sglang.srt.layers.moe.token_dispatcher.flashinfer import FlashinferDispatch
from sglang.srt.layers.moe.utils import initialize_moe_config
from sglang.srt.runtime_context import get_context, publish
from sglang.srt.server_args import ServerArgs, set_global_server_args_for_scheduler
from sglang.test.test_utils import CustomTestCase
from sglang.test.test_utils import CustomTestCase, publish_build_topology
class TestFlashinferDispatcher(CustomTestCase):
@@ -44,9 +44,8 @@ class TestFlashinferDispatcher(CustomTestCase):
publish(server_args, role="scheduler")
initialize_moe_config()
initialize_model_parallel(
tensor_model_parallel_size=world_size, expert_model_parallel_size=world_size
)
publish_build_topology(tp_size=world_size, ep_size=world_size, world_rank=rank)
initialize_model_parallel()
@classmethod
def tearDownClass(cls):