Skip to content

Fix Qwen3 MoE: also guard EP all-reduce with not use_reduce_scatter (follow-up to #23731) #23734

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
ByronHsu merged 1 commit into from
Apr 26, 2026
Merged

Fix Qwen3 MoE: also guard EP all-reduce with not use_reduce_scatter (follow-up to #23731) #23734

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
1 change: 1 addition & 0 deletions python/sglang/srt/models/qwen3_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -335,6 +335,7 @@ def forward_normal(
if (
self.ep_size > 1
and not should_allreduce_fusion
and not use_reduce_scatter
and not should_use_dp_reduce_scatterv()
Comment on lines 337 to 339
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist Bot Apr 26, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

To fully align the EP all-reduce guard with the TP all-reduce guard below (lines 344-349), you should also include not should_use_flashinfer_cutlass_moe_fp4_allgather(). This ensures that if the FlashInfer specialized communication path is active, the EP all-reduce is correctly skipped to avoid double-reduction, maintaining consistency between the TP and EP branches.

            and not should_allreduce_fusion
            and not use_reduce_scatter
            and not should_use_flashinfer_cutlass_moe_fp4_allgather()
            and not should_use_dp_reduce_scatterv()

):
final_hidden_states = moe_expert_parallel_all_reduce(final_hidden_states)
Expand Down
Loading