@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 72bac73067
Fork 0

A high-throughput and memory-efficient inference and serving engine for LLMs

Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git

Synced Nov 16, 00:45 UTC

[Build/CI] Fix libcuda.so linkage (#12424)

Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>

Tyler Michael Smith committed Jan 26, 2025 at 16:18 UTC 72bac7306796b01c202d846da041f62ded3a26a9
1 file changed +4
CMakeLists.txt
+4
@@ -446,6 +446,9 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
446 endif()
447
448 message(STATUS "Enabling C extension.")
449 +if(VLLM_GPU_LANG STREQUAL "CUDA")
450 + list(APPEND VLLM_C_LIBS cuda)
451 +endif()
452 define_gpu_extension_target(
453 _C
454 DESTINATION vllm
@@ -454,6 +457,7 @@ define_gpu_extension_target(
457 COMPILE_FLAGS ${VLLM_GPU_FLAGS}
458 ARCHITECTURES ${VLLM_GPU_ARCHES}
459 INCLUDE_DIRECTORIES ${CUTLASS_INCLUDE_DIR};${CUTLASS_TOOLS_UTIL_INCLUDE_DIR}
460 + LIBRARIES ${VLLM_C_LIBS}
461 USE_SABI 3
462 WITH_SOABI)
463