[mlir] Fix the link of libcuda.so in MLIRGPUTransforms to not use fully qualified path #74018
Conversation
|
@stellaraccident : is this the right way to do this in CMake? |
|
@llvm/pr-subscribers-mlir Author: Mehdi Amini (joker-eph) ChangesAt the moment we find libcuda.so in a path like: /usr/local/cuda/targets/x86_64-linux/lib/stubs/libcuda.so and directly add this to We're changing this to instead -L /usr/local/cuda/targets/x86_64-linux/lib/stubs/ -lcuda Full diff: https://github.com/llvm/llvm-project/pull/74018.diff 1 Files Affected:
diff --git a/mlir/lib/Dialect/GPU/CMakeLists.txt b/mlir/lib/Dialect/GPU/CMakeLists.txt
index b76d18e81246eb1..7da15974abc6565 100644
--- a/mlir/lib/Dialect/GPU/CMakeLists.txt
+++ b/mlir/lib/Dialect/GPU/CMakeLists.txt
@@ -135,11 +135,16 @@ if(MLIR_ENABLE_CUDA_RUNNER)
${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES}
)
+ # Add link path for the cuda driver library.
find_library(CUDA_DRIVER_LIBRARY cuda HINTS ${CMAKE_CUDA_IMPLICIT_LINK_DIRECTORIES} REQUIRED)
+ get_filename_component(CUDA_DRIVER_LIBRARY_PATH "${CUDA_DRIVER_LIBRARY}" DIRECTORY)
+ message("CUDA_DRIVER_LIBRARY: ${CUDA_DRIVER_LIBRARY} ${CUDA_DRIVER_LIBRARY_PATH}")
+ target_link_directories(MLIRGPUTransforms PRIVATE ${CUDA_DRIVER_LIBRARY_PATH})
+
target_link_libraries(MLIRGPUTransforms
PRIVATE
MLIRNVVMToLLVMIRTranslation
- ${CUDA_DRIVER_LIBRARY}
+ cuda
)
endif()
|
|
@llvm/pr-subscribers-mlir-gpu Author: Mehdi Amini (joker-eph) ChangesAt the moment we find libcuda.so in a path like: /usr/local/cuda/targets/x86_64-linux/lib/stubs/libcuda.so and directly add this to We're changing this to instead -L /usr/local/cuda/targets/x86_64-linux/lib/stubs/ -lcuda Full diff: https://github.com/llvm/llvm-project/pull/74018.diff 1 Files Affected:
diff --git a/mlir/lib/Dialect/GPU/CMakeLists.txt b/mlir/lib/Dialect/GPU/CMakeLists.txt
index b76d18e81246eb1..7da15974abc6565 100644
--- a/mlir/lib/Dialect/GPU/CMakeLists.txt
+++ b/mlir/lib/Dialect/GPU/CMakeLists.txt
@@ -135,11 +135,16 @@ if(MLIR_ENABLE_CUDA_RUNNER)
${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES}
)
+ # Add link path for the cuda driver library.
find_library(CUDA_DRIVER_LIBRARY cuda HINTS ${CMAKE_CUDA_IMPLICIT_LINK_DIRECTORIES} REQUIRED)
+ get_filename_component(CUDA_DRIVER_LIBRARY_PATH "${CUDA_DRIVER_LIBRARY}" DIRECTORY)
+ message("CUDA_DRIVER_LIBRARY: ${CUDA_DRIVER_LIBRARY} ${CUDA_DRIVER_LIBRARY_PATH}")
+ target_link_directories(MLIRGPUTransforms PRIVATE ${CUDA_DRIVER_LIBRARY_PATH})
+
target_link_libraries(MLIRGPUTransforms
PRIVATE
MLIRNVVMToLLVMIRTranslation
- ${CUDA_DRIVER_LIBRARY}
+ cuda
)
endif()
|
…ly qualified path At the moment we find libcuda.so in a path like: /usr/local/cuda/targets/x86_64-linux/lib/stubs/libcuda.so and directly add this to `target_link_libraries`. The problem is that our installed MLIR package will include the full path to the library, and a user downstream when including our cmake installed package will inherit this full path. We're changing this to instead -L /usr/local/cuda/targets/x86_64-linux/lib/stubs/ -lcuda
joker-eph
force-pushed
the
remove_fullyqualified_path
branch
from
a771d1c to
ee3264c
Compare
At the moment we find libcuda.so in a path like:
/usr/local/cuda/targets/x86_64-linux/lib/stubs/libcuda.so
and directly add this to
target_link_libraries. The problem is that our installed MLIR package will include the full path to the library, and a user downstream when including our cmake installed package will inherit this full path.We're changing this to instead
-L /usr/local/cuda/targets/x86_64-linux/lib/stubs/ -lcuda