Skip to content

[sgl-kernel][1/2] Fused qk_norm_rope for Qwen3-MoE #14036

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
BBuf merged 1 commit into from
Nov 28, 2025
Merged

[sgl-kernel][1/2] Fused qk_norm_rope for Qwen3-MoE #14036

Show file tree
Hide file tree
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 sgl-kernel/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -318,6 +318,7 @@ set(SOURCES
"csrc/moe/marlin_moe_wna16/ops.cu"
"csrc/moe/moe_align_kernel.cu"
"csrc/moe/moe_fused_gate.cu"
"csrc/moe/fused_qknorm_rope_kernel.cu"
"csrc/moe/kimi_k2_moe_fused_gate.cu"
"csrc/moe/moe_sum.cu"
"csrc/moe/moe_sum_reduce.cu"
Expand Down
8 changes: 8 additions & 0 deletions sgl-kernel/csrc/common_extension.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -264,9 +264,17 @@ TORCH_LIBRARY_FRAGMENT(sgl_kernel, m) {

m.def("shuffle_rows(Tensor input, Tensor dst2src_map, Tensor output) -> ()");
m.impl("shuffle_rows", torch::kCUDA, &shuffle_rows);

m.def("apply_shuffle_mul_sum(Tensor input, Tensor output, Tensor permutation, Tensor? factors) -> ()");
m.impl("apply_shuffle_mul_sum", torch::kCUDA, &apply_shuffle_mul_sum);

m.def(
"fused_qk_norm_rope(Tensor! qkv, int num_heads_q, "
"int num_heads_k, int num_heads_v, int head_dim, float eps, "
"Tensor q_weight, Tensor k_weight, float base, "
"bool is_neox, Tensor position_ids, float factor, float low, float high, float attention_factor) -> ()");
m.impl("fused_qk_norm_rope", torch::kCUDA, &fused_qk_norm_rope);

/*
* From csrc/moe/cutlass_moe/w4a8
*/
Expand Down
Loading
Loading