SotaMunger opened a new issue, #48244: URL: https://github.com/apache/arrow/issues/48244
### Describe the bug, including details regarding any error messages, version, and platform. Hi all, I'm trying to install pyarrow with CUDA support in a conda environment according to the instructions detailed here: https://arrow.apache.org/docs/developers/python/building.html. The cmake --build arrow/cpp/build --target install step fails and gives the following output: Building CXX object src/arrow...iles/arrow-cuda-test.dir/cuda_test.cc.o FAILED: [code=1] src/arrow/gpu/CMakeFiles/arrow-cuda-test.dir/cuda_test.cc.o /home/richard/miniconda3/envs/money/bin/ccache /home/richard/miniconda3/envs/money/bin/x86_64-conda-linux-gnu-c++ -DARROW_HAVE_RUNTIME_AVX2 -DARROW_HAVE_RUNTIME_AVX512 -DARROW_HAVE_RUNTIME_BMI2 -DARROW_HAVE_RUNTIME_SSE4_2 -DARROW_HAVE_SSE4_2 -DARROW_WITH_TIMING_TESTS -DGTEST_LINKED_AS_SHARED_LIBRARY=1 -I/home/richard/arrow/cpp/build/src -I/home/richard/arrow/cpp/src -I/home/richard/arrow/cpp/src/generated -isystem /home/richard/arrow/cpp/thirdparty/flatbuffers/include -isystem /usr/local/cuda/targets/x86_64-linux/include -isystem /usr/local/cuda/targets/x86_64-linux/include/cccl -Wno-noexcept-type -Wno-self-move -fvisibility-inlines-hidden -fmessage-length=0 -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /home/richard/miniconda3/envs/money/include -fdiagnostics-color=always -Wall -fno-semantic-interposition -msse4.2 -O3 -DNDEBUG -O2 -ftree-vectorize -std=c++17 -fPIE -MD -MT src/arrow/gpu/CMakeFiles/ar row-cuda-test.dir/cuda_test.cc.o -MF src/arrow/gpu/CMakeFiles/arrow-cuda-test.dir/cuda_test.cc.o.d -o src/arrow/gpu/CMakeFiles/arrow-cuda-test.dir/cuda_test.cc.o -c /home/richard/arrow/cpp/src/arrow/gpu/cuda_test.cc In file included from /home/richard/arrow/cpp/src/arrow/result.h:27, from /home/richard/arrow/cpp/src/arrow/device.h:26, from /home/richard/arrow/cpp/src/arrow/c/bridge.h:25, from /home/richard/arrow/cpp/src/arrow/gpu/cuda_test.cc:26: /home/richard/arrow/cpp/src/arrow/gpu/cuda_test.cc: In member function 'arrow::Result<CUctx_st*> arrow::cuda::TestCudaBase::NonPrimaryRawContext()': /home/richard/arrow/cpp/src/arrow/gpu/cuda_test.cc:94:45: error: too few arguments to function 'CUresult cuCtxCreate_v4(CUctx_st**, CUctxCreateParams*, unsigned int, CUdevice)' 94 | RETURN_NOT_OK(StatusFromCuda(cuCtxCreate(&ctx, /*flags=*/0, device_->handle()))); | ^ /home/richard/arrow/cpp/src/arrow/status.h:58:45: note: in definition of macro 'ARROW_RETURN_NOT_OK' 58 | ::arrow::Status __s = ::arrow::ToStatus(status); \ | ^~~~~~ /home/richard/arrow/cpp/src/arrow/gpu/cuda_test.cc:94:5: note: in expansion of macro 'RETURN_NOT_OK' 94 | RETURN_NOT_OK(StatusFromCuda(cuCtxCreate(&ctx, /*flags=*/0, device_->handle()))); | ^~~~~~~~~~~~~ In file included from /home/richard/arrow/cpp/src/arrow/gpu/cuda_test.cc:22: /usr/local/cuda/targets/x86_64-linux/include/cuda.h:6275:18: note: declared here 6275 | CUresult CUDAAPI cuCtxCreate(CUcontext *pctx, CUctxCreateParams *ctxCreateParams, unsigned int flags, CUdevice dev); | ^~~~~~~~~~~ [69/273] Building CXX object src/parqu...writer-test.dir/column_writer_test.cc.o ninja: build stopped: subcommand failed. I'm running Ubuntu 24.04.3 LTS and have cuda toolkit 13.0 and the NVIDIA 580 open driver installed. Could anybody explain to me how to resolve the error? Thanks. ### Component(s) Python -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
