mirrors/llama.cpp
0
0 Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-10-28 09:45:55 +08:00
Code Activity

Explicitly link against cublaslt

Browse Source
This commit is contained in:
Oliver Simons committed 2026-10-27 18:47:47 +01:00
1 parent d95220fb33
commit be52613450
1 file changed
+3 -2
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/CMakeLists.txt
+3 -2
View File
@@ -160,10 +160,11 @@ if (CUDAToolkit_FOUND)
target_link_libraries(ggml-cuda PRIVATE CCCL::CCCL)
endif()
# cuBLASLt is a transitive dependency of cuBLAS, but FP8 matmuls call its API directly.
if (GGML_STATIC)
if (WIN32)
# As of 12.3.1 CUDA Toolkit for Windows does not offer a static cublas library
target_link_libraries(ggml-cuda PRIVATE CUDA::cudart_static CUDA::cublas)
target_link_libraries(ggml-cuda PRIVATE CUDA::cudart_static CUDA::cublas CUDA::cublasLt)
else ()
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL "10.1")
target_link_libraries(ggml-cuda PRIVATE CUDA::cudart_static CUDA::cublas_static CUDA::cublasLt_static)
@@ -172,7 +173,7 @@ if (CUDAToolkit_FOUND)
endif()
endif()
else()
target_link_libraries(ggml-cuda PRIVATE CUDA::cudart CUDA::cublas)
target_link_libraries(ggml-cuda PRIVATE CUDA::cudart CUDA::cublas CUDA::cublasLt)
endif()
if (GGML_CUDA_NO_VMM)