Skip to content

Navigation Menu

Sign in
Sign up

Avoid building cuvs in pip devcontainer builds #8520

Open
Assignees

Description

Per @bdice

If cuml is building cuvs from source in its devcontainers, we can probably change that and install packages nowadays.

In pip devcontainer builds we're seeing cuvs source being built. I tried to investigate why, but couldn't figure it out through the dependencies.yaml/cmake/... stack.

Example build log: https://github.com/NVIDIA/cuml/actions/runs/32778840561/job/97596127724

...
[10/5494] /usr/bin/sccache /usr/local/cuda/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/usr/bin/g++ -DCCCL_DISABLE_PDL -DCUB_DISABLE_NAMESPACE_MAGIC -DCUB_IGNORE_NAMESPACE_MAGIC_ERROR -DRAFT_LOG_ACTIVE_LEVEL=RAPIDS_LOGGER_LOG_LEVEL_INFO -DRAFT_SYSTEM_LITTLE_ENDIAN=1 -DTHRUST_DEVICE_SYSTEM=THRUST_DEVICE_SYSTEM_CUDA -DTHRUST_DISABLE_ABI_NAMESPACE -DTHRUST_HOST_SYSTEM=THRUST_HOST_SYSTEM_CPP -DTHRUST_IGNORE_ABI_NAMESPACE_ERROR -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cuvs-src/cpp/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cuvs-src/cpp/src -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cuvs-src/cpp/../c/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/rmm-src/cpp/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/rmm-build/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cccl-src/lib/cmake/thrust/../../../thrust -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cccl-src/lib/cmake/libcudacxx/../../../libcudacxx/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cccl-src/lib/cmake/cub/../../../cub -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/nvtx3-src/c/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/rapids_logger-src/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/raft-src/cpp/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/raft-build/include -I/home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cuco-src/include -isystem /usr/local/cuda/targets/x86_64-linux/include -isystem /usr/local/cuda/targets/x86_64-linux/include/cccl -t=7 -O3 -DNDEBUG -std=c++20 "--generate-code=arch=compute_75,code=[lto_75]" -Xcompiler=-fPIC -Xfatbin=--compress-all --compress-mode=size -Xcompiler=-Wno-deprecated-declarations -DRAFT_HIDE_DEPRECATION_WARNINGS -Xcompiler=-Wall,-Werror,-Wno-error=deprecated-declarations,-Wno-reorder -Werror=all-warnings --expt-extended-lambda --expt-relaxed-constexpr -DCUDA_API_PER_THREAD_DEFAULT_STREAM -Xfatbin=-compress-all -Xcompiler=-fopenmp -MD -MT _deps/cuvs-build/CMakeFiles/ivf_flat_interleaved_scan_capacity_128_ascending_data_i8_acc_i32_index_i64_kernel.dir/generated_kernels/ivf_flat/interleaved_scan/ivf_flat_interleaved_scan_capacity_128_ascending_data_i8_acc_i32_index_i64_kernel.fatbin -MF _deps/cuvs-build/CMakeFiles/ivf_flat_interleaved_scan_capacity_128_ascending_data_i8_acc_i32_index_i64_kernel.dir/generated_kernels/ivf_flat/interleaved_scan/ivf_flat_interleaved_scan_capacity_128_ascending_data_i8_acc_i32_index_i64_kernel.fatbin.d -x cu -rdc=true -fatbin /home/coder/cuml/cpp/build/pip/cuda-13.3/release/_deps/cuvs-build/generated_kernels/ivf_flat/interleaved_scan/ivf_flat_interleaved_scan_capacity_128_ascending_data_i8_acc_i32_index_i64_kernel.cu -o _deps/cuvs-build/CMakeFiles/ivf_flat_interleaved_scan_capacity_128_ascending_data_i8_acc_i32_index_i64_kernel.dir/generated_kernels/ivf_flat/interleaved_scan/ivf_flat_interleaved_scan_capacity_128_ascending_data_i8_acc_i32_index_i64_kernel.fatbin
...

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Labels

No labels
No labels

Type

Projects

No projects

Milestone

No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions

    AltStyle によって変換されたページ (->オリジナル) /