Skip to content

Qwen3next flashinfer allreduce auto enable #22664

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

Open
BBuf wants to merge 2 commits into
base: main
Choose a base branch
from
+3 −1
Open

Qwen3next flashinfer allreduce auto enable #22664

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7b95e92
server: auto-enable flashinfer allreduce for qwen3next
BBuf Apr 13, 2026
f2176f6
Drop Qwen3 Next server args test coverage from PR
BBuf Apr 13, 2026
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
4 changes: 3 additions & 1 deletion python/sglang/srt/server_args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2171,7 +2171,8 @@ def _handle_model_specific_adjustments(self):
)

# TRTLLM AllReduce Fusion supports SM90/100, enable it by default
# for models with explicit support (DeepseekV3, GptOss, Glm4Moe, Qwen3Moe)
# for models with explicit support (DeepseekV3, GptOss, Glm4Moe,
# Qwen3/Qwen3Next/Qwen3.5 MoE families)
# TODO: currently, it is only supported in the single node scenario. https://github.com/flashinfer-ai/flashinfer/issues/2006
# TODO: there is currently a bug on H20 device specifically, https://github.com/flashinfer-ai/flashinfer/issues/2204
device_name = get_device_name()
Expand All @@ -2189,6 +2190,7 @@ def _handle_model_specific_adjustments(self):
"Glm4MoeForCausalLM",
"Glm4MoeLiteForCausalLM",
"Qwen3MoeForCausalLM",
"Qwen3NextForCausalLM",
"KimiK25ForConditionalGeneration",
"Qwen3_5MoeForConditionalGeneration",
"Qwen3_5ForConditionalGeneration",
Expand Down
Loading