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

[sgl-kernel slimming] remove sgl-kernel moe-wna16-marlin (#19379)

Browse Source
This commit is contained in:
Xiaoyu Zhang
2026-02-27 09:17:46 +08:00
committed by GitHub
parent 5194eef88b
commit 054bd71086
16 changed files with 0 additions and 3721 deletions
Download Patch File Download Diff File Expand all files Collapse all files
sgl-kernel/csrc/common_extension.cc
-17
View File
@@ -287,23 +287,6 @@ TORCH_LIBRARY_FRAGMENT(sgl_kernel, m) {
" int chunk_size, int topk) -> ()");
m.impl("cutlass_w4a8_moe_mm", torch::kCUDA, &cutlass_w4a8_moe_mm);
/*
* From csrc/moe/marlin_moe_wna16
*/
m.def(
"moe_wna16_marlin_gemm(Tensor! a, Tensor? c_or_none,"
"Tensor! b_q_weight, Tensor? b_bias_or_none, Tensor! b_scales,"
"Tensor? global_scale_or_none, Tensor? b_zeros_or_none,"
"Tensor? g_idx_or_none, Tensor? perm_or_none, Tensor! workspace,"
"Tensor sorted_token_ids,"
"Tensor! expert_ids, Tensor! num_tokens_past_padded,"
"Tensor! topk_weights, int moe_block_size, int top_k, "
"bool mul_topk_weights, bool is_ep, int b_q_type_id,"
"int size_m, int size_n, int size_k,"
"bool is_k_full, bool use_atomic_add,"
"bool use_fp32_reduce, bool is_zp_float) -> Tensor");
m.impl("moe_wna16_marlin_gemm", torch::kCUDA, &moe_wna16_marlin_gemm);
/*
* From csrc/speculative
*/