EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

support cmake for sgl-kernel (#4706)

Browse Source 
Co-authored-by: hebiao064 <hebiaobuaa@gmail.com>
Co-authored-by: yinfan98 <1106310035@qq.com>
This commit is contained in:
Yineng Zhang
2025-03-27 01:42:28 -07:00
committed by GitHub
parent 1b9175cb23
commit 8bf6d7f406
18 changed files with 426 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
sgl-kernel/csrc/attention/lightning_attention_decode_kernel.cu
View File

@@ -18,7 +18,7 @@ limitations under the License.
#include <c10/cuda/CUDAGuard.h>
#include <cuda_fp16.h>
#include <cuda_runtime.h>
#include <torch/extension.h>
#include <torch/all.h>
#define THREADS_PER_BLOCK 128

1
sgl-kernel/csrc/gemm/cublas_grouped_gemm.cu
View File

@@ -12,7 +12,6 @@
#include <cuda_fp16.h>
#include <cuda_runtime.h>
#include <torch/all.h>
#include <torch/extension.h>
#include <cstdio>
#include <cstdlib>

1
sgl-kernel/csrc/moe/moe_align_kernel.cu
View File

@@ -16,7 +16,6 @@ limitations under the License.
#include <ATen/ATen.h>
#include <ATen/cuda/CUDAContext.h>
#include <c10/cuda/CUDAGuard.h>
#include <torch/extension.h>
#include <THC/THCAtomics.cuh>

6
sgl-kernel/csrc/torch_extension.cc
View File

@@ -12,8 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/
#include <ATen/core/dispatch/Dispatcher.h>
#include <torch/all.h>
#include <torch/library.h>
#include "sgl_kernel_ops.h"
@@ -178,9 +178,9 @@ TORCH_LIBRARY_EXPAND(sgl_kernel, m) {
m.impl("min_p_sampling_from_probs", torch::kCUDA, &min_p_sampling_from_probs);
m.def(
"top_k_renorm_probs_wrapper(Tensor probs, Tensor! renorm_probs, Tensor? maybe_top_k_arr, int top_k_val, int "
"top_k_renorm_probs(Tensor probs, Tensor! renorm_probs, Tensor? maybe_top_k_arr, int top_k_val, int "
"cuda_stream) -> ()");
m.impl("top_k_renorm_probs_wrapper", torch::kCUDA, &top_k_renorm_probs_wrapper);
m.impl("top_k_renorm_probs", torch::kCUDA, &top_k_renorm_probs);
m.def(
"top_p_renorm_probs(Tensor probs, Tensor! renorm_probs, Tensor? maybe_top_p_arr, float top_p_val, int "