xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 15:25:28 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CI/Build] Update checking logic in cutlass_group_gemm_supported (#27948)

Browse Source 
Signed-off-by: zhewenli <zhewenli@meta.com>
This commit is contained in:
Zhewen Li 2025-11-05 15:40:10 -08:00 committed by GitHub
parent e15601789b
commit 5ee93a5956
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/_custom_ops.py
View File

@ -836,7 +836,11 @@ def cutlass_sparse_scaled_mm_supported(cuda_device_capability: int) -> bool:
def cutlass_group_gemm_supported(cuda_device_capability: int) -> bool: def cutlass_group_gemm_supported(cuda_device_capability: int) -> bool:
return torch.ops._C.cutlass_group_gemm_supported(cuda_device_capability) try:
return torch.ops._C.cutlass_group_gemm_supported(cuda_device_capability)
except AttributeError:
# Return False on non-CUDA platforms where it is not available
return False
def cutlass_sparse_compress(a: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]: def cutlass_sparse_compress(a: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]: