Skip to content

[https://nvbugs/5467531][fix] Unwaive fused_moe all to all test with DeepEPLowLatency #9617

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
liji-nv merged 3 commits into from
Dec 4, 2025
Merged

[https://nvbugs/5467531][fix] Unwaive fused_moe all to all test with DeepEPLowLatency #9617

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
75d0113
[https://nvbugs/5467531][fix] Unwaive fused_moe all to all test with …
liji-nv Dec 2, 2025
610d74f
Fix MI merge conflict
liji-nv Dec 2, 2025
2544475
Fix another conflict on shape
liji-nv Dec 3, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -966,6 +966,7 @@ def forward_fake(
output_dtype: Optional[torch.dtype] = None,
all_rank_num_tokens: Optional[List[int]] = None,
use_dp_padding: Optional[bool] = None,
alltoall_result_do_sum: bool = True,
**kwargs,
) -> Union[torch.Tensor, List[torch.Tensor]]:
moe_output = super().forward_fake(
Expand All @@ -976,7 +977,7 @@ def forward_fake(
all_rank_num_tokens=all_rank_num_tokens,
use_dp_padding=use_dp_padding,
**kwargs)
if self.alltoall_method_type == AlltoallMethodType.MNNVL:
if self.alltoall_method_type == AlltoallMethodType.NVLinkTwoSided and not alltoall_result_do_sum:
shape = moe_output.shape
top_k = self.routing_method.experts_per_token
new_shape = [shape[0], top_k, shape[1]]
Expand Down
4 changes: 0 additions & 4 deletions tests/unittest/_torch/modules/test_fused_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -335,10 +335,6 @@ def per_rank_test_fused_moe_alltoall(job_id):
],
ids=lambda s: s.name)
def test_fused_moe_alltoall_fp4(alltoall_method_type):

if alltoall_method_type == AlltoallMethodType.DeepEPLowLatency:
pytest.skip("Skipped due to https://nvbugs/5467531")

world_size = 4
dtype = torch.bfloat16
HIDDEN_SIZE = 4096
Expand Down