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

[Kernel Slimming] Remove sgl-kernel AOT marlin kernels (#19241)

Browse Source
This commit is contained in:
Xiaoyu Zhang
2026-02-25 10:08:22 +08:00
committed by GitHub
parent 3b89302277
commit 9dff933164
13 changed files with 0 additions and 2136 deletions
Download Patch File Download Diff File Expand all files Collapse all files
sgl-kernel/csrc/common_extension.cc
-14
View File
@@ -196,14 +196,6 @@ TORCH_LIBRARY_FRAGMENT(sgl_kernel, m) {
/*
* From csrc/gemm/gptq
*/
m.def(
"gptq_marlin_gemm(Tensor! a, Tensor? c_or_none,"
"Tensor! b_q_weight, Tensor! b_scales, Tensor? global_scale_or_none,"
"Tensor? b_zeros_or_none, Tensor? g_idx_or_none, Tensor? perm_or_none,"
"Tensor! workspace, 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("gptq_marlin_gemm", torch::kCUDA, &gptq_marlin_gemm);
m.def(
"gptq_gemm(Tensor a, Tensor b_q_weight, Tensor b_gptq_qzeros, Tensor b_gptq_scales, Tensor b_g_idx, bool "
"use_shuffle, int bit) -> Tensor");
@@ -212,12 +204,6 @@ TORCH_LIBRARY_FRAGMENT(sgl_kernel, m) {
m.def("gptq_shuffle(Tensor! q_weight, Tensor q_perm, int bit) -> ()");
m.impl("gptq_shuffle", torch::kCUDA, &gptq_shuffle);
m.def("gptq_marlin_repack(Tensor! b_q_weight, Tensor! perm, int size_k, int size_n, int num_bits) -> Tensor");
m.impl("gptq_marlin_repack", torch::kCUDA, &gptq_marlin_repack);
m.def("awq_marlin_repack(Tensor! b_q_weight, int size_k, int size_n, int num_bits) -> Tensor");
m.impl("awq_marlin_repack", torch::kCUDA, &awq_marlin_repack);
/*
* From csrc/moe
*/