xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 00:06:06 +08:00
Code Issues Packages Projects Releases Wiki Activity

Remove sm120 arch from sm100 cutlass kernel arch list (#19716)

Browse Source 
Signed-off-by: mgoin <mgoin64@gmail.com>
This commit is contained in:
Michael Goin 2025-06-18 03:49:39 +09:00 committed by GitHub
parent ffb2cd6b54
commit bf57ccc5c2
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
CMakeLists.txt
View File

@ -420,9 +420,9 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
endif() endif()
endif() endif()
# The cutlass_scaled_mm kernels for Blackwell (c3x, i.e. CUTLASS 3.x) require # The cutlass_scaled_mm kernels for Blackwell SM100 (c3x, i.e. CUTLASS 3.x)
# CUDA 12.8 or later # require CUDA 12.8 or later
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a;12.0a" "${CUDA_ARCHS}") cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND SCALED_MM_ARCHS) if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND SCALED_MM_ARCHS)
set(SRCS set(SRCS
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm100.cu" "csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm100.cu"