83 lines
3.3 KiB
Docker
83 lines
3.3 KiB
Docker
ARG CUDA_VERSION_ARG
|
|
FROM nvcr.io/nvidia/cuda:$CUDA_VERSION_ARG-devel-rockylinux8
|
|
ARG CUDA_VERSION_ARG
|
|
ARG NCCL_VERSION_ARG
|
|
ARG RAPIDS_VERSION_ARG
|
|
|
|
# Install all basic requirements
|
|
RUN \
|
|
curl -fsSL https://developer.download.nvidia.com/compute/cuda/repos/rhel7/x86_64/D42D0685.pub | sed '/^Version/d' \
|
|
> /etc/pki/rpm-gpg/RPM-GPG-KEY-NVIDIA && \
|
|
dnf -y update && \
|
|
dnf -y install dnf-plugins-core && \
|
|
dnf config-manager --set-enabled powertools && \
|
|
dnf install -y tar unzip wget xz git which ninja-build gcc-toolset-10-gcc gcc-toolset-10-binutils gcc-toolset-10-gcc-c++ && \
|
|
# Python
|
|
wget -nv -O conda.sh https://github.com/conda-forge/miniforge/releases/download/24.3.0-0/Miniforge3-24.3.0-0-Linux-x86_64.sh && \
|
|
bash conda.sh -b -p /opt/miniforge && \
|
|
/opt/miniforge/bin/python -m pip install awscli && \
|
|
# CMake
|
|
wget -nv -nc https://cmake.org/files/v3.29/cmake-3.29.5-linux-x86_64.sh --no-check-certificate && \
|
|
bash cmake-3.29.5-linux-x86_64.sh --skip-license --prefix=/usr
|
|
|
|
# NCCL2 (License: https://docs.nvidia.com/deeplearning/sdk/nccl-sla/index.html)
|
|
RUN \
|
|
export CUDA_SHORT=`echo $CUDA_VERSION_ARG | grep -o -E '[0-9]+\.[0-9]'` && \
|
|
export NCCL_VERSION=$NCCL_VERSION_ARG && \
|
|
dnf config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/rhel7/x86_64/cuda-rhel7.repo && \
|
|
dnf -y update && \
|
|
dnf install -y libnccl-${NCCL_VERSION}+cuda${CUDA_SHORT} libnccl-devel-${NCCL_VERSION}+cuda${CUDA_SHORT}
|
|
|
|
ENV PATH=/opt/miniforge/bin:/usr/local/ninja:$PATH
|
|
ENV CC=/opt/rh/gcc-toolset-10/root/usr/bin/gcc
|
|
ENV CXX=/opt/rh/gcc-toolset-10/root/usr/bin/c++
|
|
ENV CPP=/opt/rh/gcc-toolset-10/root/usr/bin/cpp
|
|
ENV CUDAHOSTCXX=/opt/rh/gcc-toolset-10/root/usr/bin/c++
|
|
|
|
ENV GOSU_VERSION=1.10
|
|
|
|
# Install gRPC
|
|
# Patch Abseil to apply https://github.com/abseil/abseil-cpp/issues/1629
|
|
RUN git clone -b v1.65.4 https://github.com/grpc/grpc.git \
|
|
--recurse-submodules --depth 1 && \
|
|
pushd grpc && \
|
|
pushd third_party/abseil-cpp && \
|
|
git fetch origin master && \
|
|
git cherry-pick -n cfde5f74e276049727f9556f13473a59fe77d9eb && \
|
|
popd && \
|
|
cmake -S . -B build -GNinja -DCMAKE_INSTALL_PREFIX=/opt/grpc -DCMAKE_CXX_VISIBILITY_PRESET=hidden && \
|
|
cmake --build build --target install && \
|
|
popd && \
|
|
rm -rf grpc
|
|
|
|
# Install RMM
|
|
# Patch out -Werror
|
|
# Patch CCCL 2.5.0 to apply https://github.com/NVIDIA/cccl/pull/1957
|
|
RUN git clone -b v${RAPIDS_VERSION_ARG}.00 https://github.com/rapidsai/rmm.git --recurse-submodules --depth 1 && \
|
|
pushd rmm && \
|
|
find . -name CMakeLists.txt -print0 | xargs -0 sed -i 's/-Werror//g' && \
|
|
mkdir build && \
|
|
pushd build && \
|
|
cmake .. -GNinja -DCMAKE_INSTALL_PREFIX=/opt/rmm -DCUDA_STATIC_RUNTIME=ON && \
|
|
pushd _deps/cccl-src/ && \
|
|
git fetch origin main && \
|
|
git cherry-pick -n 9fcb32c228865f21f2b002b29d38a06b4c6fbd73 && \
|
|
popd && \
|
|
cmake --build . --target install && \
|
|
popd && \
|
|
popd && \
|
|
rm -rf rmm
|
|
|
|
# Install lightweight sudo (not bound to TTY)
|
|
RUN set -ex; \
|
|
wget -nv -nc -O /usr/local/bin/gosu "https://github.com/tianon/gosu/releases/download/$GOSU_VERSION/gosu-amd64" && \
|
|
chmod +x /usr/local/bin/gosu && \
|
|
gosu nobody true
|
|
|
|
# Default entry-point to use if running locally
|
|
# It will preserve attributes of created files
|
|
COPY entrypoint.sh /scripts/
|
|
|
|
WORKDIR /workspace
|
|
ENTRYPOINT ["/scripts/entrypoint.sh"]
|