Dockerfile-build

ARG docker_name

FROM $docker_name

MAINTAINER H2o.ai <ops@h2o.ai>

#
# Env variables for CUDA. Necessary because certain systems don't support nvidia-docker so we should use plain docker as much as possible.
#
ENV HOME=/root
ENV CUDA_HOME=/usr/local/cuda
ENV CUDADIR=/usr/local/cuda/include/
ENV PATH=/usr/local/cuda/bin:$PATH
ENV LD_LIBRARY_PATH_CUDA=$CUDA_HOME/lib64/:$CUDA_HOME/lib/:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64
ENV LD_LIBRARY_PATH_BUILD=/lib64:/usr/local/lib64:/home/$USER/lib/
ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH_CUDA:$LD_LIBRARY_PATH_BUILD:$LD_LIBRARY_PATH

#
# Env variables used by the codebase.
#
ENV OMP_NUM_THREADS=32
ENV MKL_NUM_THREADS=32
ENV VECLIB_MAXIMUM_THREADS=32
ENV GIT_VERSION=2.17.0

#
# Install necessary libraries and dependencies
#
RUN yum install -y epel-release

# RUN yum update -y

RUN yum install centos-release-scl-rh -y

# Setup gcc etc.
RUN yum install -y gcc gcc-c++ libgcc libstdc++ libgomp glibc

# Git & gcc requirements
RUN yum install -y libcurl-devel zlib-devel asciidoc xmlto wget make autoconf gettext gmp-devel mpfr-devel libmpc-devel

RUN yum install -y devtoolset-7

# Compile from source because yum's latest version is 1.8.3
# --depth for submodule update which we use was added in 1.8.4
RUN \
    wget https://www.kernel.org/pub/software/scm/git/git-${GIT_VERSION}.tar.xz && \
    tar xf git-${GIT_VERSION}.tar.xz && \
    cd git-${GIT_VERSION} && \
    make configure && \
    ./configure --prefix=/usr && \
    make all && \
    make install;

#H2O4GPU requirements
RUN yum install -y \
    ncurses-devel \
    bzip2 \
    which \
    axel \
    openssl-devel \
    libpng-devel \
    freetype-devel \
    blas-devel \
    epel-release \
    zeromq-devel \
    openblas-devel \
    libffi-devel

RUN \
    git clone https://github.com/NVIDIA/nccl.git && \
    cd nccl && \
    git checkout tags/v2.4.7-1 && \
    scl enable devtoolset-7 'make CUDA8_GENCODE="-gencode=arch=compute_35,code=sm_35 -gencode=arch=compute_50,code=sm_50 -gencode=arch=compute_60,code=sm_60 -gencode=arch=compute_61,code=sm_61" -j src.build';

# cmake 3.17.1 looks incompatible with CUDA
RUN wget https://github.com/Kitware/CMake/releases/download/v3.16.8/cmake-3.16.8.tar.gz && \
    tar -zxvf cmake-3.16.8.tar.gz && \
    cd cmake-3.16.8 && \
    ./bootstrap && \
    make -j`nproc` && \
    make install

RUN mkdir -p /opt/h2oai/h2o4gpu

RUN wget https://s3.amazonaws.com/artifacts.h2o.ai/releases/ai/h2o/dai-thirdparty-deps/1.0-master-22/`arch`-centos7/llvm.tar.bz2 && \
    tar xvf llvm.tar.bz2 && \
    cp -r llvm/* /opt/h2oai/h2o4gpu/ && \
    rm -rf llvm*

ENV LLVM4=/opt/h2oai/h2o4gpu
ENV PATH=$LLVM4/bin:$PATH
ENV LD_LIBRARY_PATH=$LLVM4/lib:$LD_LIBRARY_PATH
ENV LLVM_CONFIG=$LLVM4/bin/llvm-config

#
# Library versions
#
ARG python_version
ENV MINICONDA_VERSION=4.8.3
ENV SWIG_VERSION=3.0.12

# conda
RUN wget https://repo.anaconda.com/miniconda/Miniconda3-py38_${MINICONDA_VERSION}-Linux-`arch`.sh && \
    bash Miniconda3-py38_${MINICONDA_VERSION}-Linux-`arch`.sh -b -p /opt/h2oai/h2o4gpu/python && \
    /opt/h2oai/h2o4gpu/python/bin/conda install -y python=${python_version} conda-build pip && \
    /opt/h2oai/h2o4gpu/python/bin/conda update conda-build


ENV PATH=/opt/h2oai/h2o4gpu/python/bin:$PATH
ENV PATH=/usr/local/bin:$PATH

#
# Symlinks
#

# AR for conda
RUN ln /usr/bin/ar $LLVM4/bin/`arch`-conda_cos6-linux-gnu-ar

# Symlinks for Python libs used by SWIG in CMake - it does not recognize Miniconda paths otherwise
RUN \
    mkdir -p /usr/lib64/ && \
    ln -sf /opt/h2oai/h2o4gpu/python/lib/libpython* /usr/lib64/ && \
    mkdir -p /usr/include/python${python_version}m && \
    ln -s /opt/h2oai/h2o4gpu/python/include/python${python_version}m/* /usr/include/python${python_version}m

# Yumming openblas puts some files in a not-so-standard locations
RUN ln -s /usr/include/openblas/* /usr/local/include/

# Symlinks for NVML
RUN \
    mkdir -p /usr/lib64/nvidia/ && \
    ln -s /usr/local/cuda-`nvcc --version | tail -n 1 | cut -f 5 -d' ' | cut -f 1 -d ','`/targets/`arch`-linux/lib/stubs/libnvidia-ml.so /usr/lib64/nvidia/libnvidia-ml.so

#
# Builds from source due to too old versions in yum
#
WORKDIR $HOME

# SWIG
RUN \
    wget https://0xdata-public.s3.amazonaws.com/swig/swig-${SWIG_VERSION}.tar.gz && \
    tar -zxvf swig-${SWIG_VERSION}.tar.gz && \
    cd swig-${SWIG_VERSION} && \
    ./configure --prefix=/usr && \
    make -j $(nproc) && \
    make install && \
    cd $HOME && \
    rm -rf swig-3*

#
# PPC64 specific - certain libs/whl don't support PPC64LE
#

# Arrow
# Need to enable c99 manually https://github.com/numpy/numpy/issues/14147
RUN bash -c 'if [ `arch` = "ppc64le" ]; then \
    git clone https://github.com/apache/arrow.git && \
    cd $HOME/arrow/cpp && \
    git checkout tags/apache-arrow-0.17.1 && \
    yum install -y boost-devel && \
    CFLAGS=-std=c99 pip install numpy==1.19.2 cython==0.29.14 scipy==1.5.2 && \
    cmake -DARROW_CXXFLAGS="-lutil" -DARROW_PYTHON=on && make -j && make install && \
    cd $HOME/arrow/python && \
    ARROW_HOME=/usr/local python setup.py install && \
    yum install -y libjpeg-devel;\
    fi'


#
# Install Python requirements
#
RUN echo ${python_version}

RUN pip install -U pip==23.3.2

RUN pip install numpy==1.23.5 scipy==1.10.1 setuptools==58.5.3

COPY src/interface_py/requirements_buildonly.txt requirements_buildonly.txt

RUN pip install -r requirements_buildonly.txt

RUN mkdir -p /etc/OpenCL/vendors && \
    echo "libnvidia-opencl.so.1" > /etc/OpenCL/vendors/nvidia.icd

RUN (localedef -v -c -i en_US -f UTF-8 en_US.UTF-8 || true)

ENV LANG en_US.UTF-8
ENV LANGUAGE en_US:en


# See https://github.com/Microsoft/LightGBM/wiki/Installation-Guide#with-gpu-support for details
# https://github.com/Microsoft/LightGBM/pull/929/files
# Could compile with these as well: -DBOOST_COMPUTE_USE_OFFLINE_CACHE=OFF -DBOOST_COMPUTE_THREAD_SAFE=ON
RUN \
    export CUDA_HOME=/usr/local/cuda/ && \
    yum install -y opencl-headers icu libicu-devel bzip2 bzip2-devel zlib-devel python-devel && \
    wget https://github.com/boostorg/boost/releases/download/boost-1.82.0/boost-1.82.0.tar.gz  && \
    tar -zxvf boost-1.82.0.tar.gz && \
    cd boost-1.82.0 && \
    export PYTHONPATH=/opt/h2oai/h2o4gpu/python/ && \
    ./bootstrap.sh --prefix=/opt/boost/ --with-python=python3 && \
    export CPPFLAGS="-I/opt/h2oai/h2o4gpu/python/include/python${python_version}/ -fPIC" && \
    export C_INCLUDE_PATH="/opt/h2oai/h2o4gpu/python/include/python${python_version}/" && export CPLUS_INCLUDE_PATH="/opt/h2oai/h2o4gpu/python/include/python${python_version}m/" && \
    ./b2 link=static -a -d0 install --prefix=/opt/boost/ --with=all -j 20 cxxflags="-fPIC -I /opt/h2oai/h2o4gpu/python/include/python${python_version}/" && \
    cd /usr/include && rm -rf boost && ln -s /opt/boost/include/boost . && \
    cd /usr/lib64/ && rm -rf libboost* && cp -a /opt/boost/lib/* . && \
    cd / 

#	yum install -y boost boost-devel boost-system boost-filesystem boost-thread

ENV LD_LIBRARY_PATH=/opt/boost/lib/:$LD_LIBRARY_PATH

RUN chmod -R o+rwx /opt/h2oai/h2o4gpu/python
RUN chmod -R o+rwx /root
RUN yum install -y hdf5-devel
RUN bash -c 'if [ `arch` == "ppc64le" ]; then \
    yum install -y ocl-icd; \
    fi'

ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/lib64/nvidia
ENV CUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda

WORKDIR $HOME

ENV GIT_AUTHOR_NAME="anonymous"
ENV GIT_AUTHOR_EMAIL="anonymous@h2o.ai"
ENV GIT_COMMITTER_NAME="anonymous"
ENV GIT_COMMITTER_EMAIL="anonymous@h2o.ai"
ENV EMAIL="anonymous@h2o.ai"