Dockerfile-runtime-single-gpu

ARG docker_name

FROM $docker_name

ARG use_miniconda

MAINTAINER H2o.ai <ops@h2o.ai>

ENV HOME=/root
ENV CUDA_HOME=/usr/local/cuda
ENV PATH=/usr/local/cuda/bin:$PATH
ENV CUDADIR=/usr/local/cuda/include/
ENV LD_LIBRARY_PATH=/usr/lib64:/usr/local/lib:$LD_LIBRARY_PATH

ENV MINICONDA_VERSION=4.8.3

# Setup gcc etc.
RUN yum install -y epel-release

RUN yum install -y gcc gcc-c++ libgcc libstdc++ libgomp glibc

RUN yum install -y \
    make \
    ncurses-devel \
    zlib-devel \
    epel-release \
    zeromq-devel \
    wget \
    blas-devel \
    openblas-devel \
    libpng-devel \
    freetype-devel \
    bzip2 \
    libffi-devel

# Arrow
RUN bash -c 'if [ `arch` = "ppc64le" ]; then yum install -y git boost-devel cmake3  autoconf libjpeg-devel; fi'

RUN mkdir -p /opt/h2oai/h2o4gpu


ARG python_version

# ln -sf /usr/bin/python36 /usr/bin/python breaks yum
RUN if [ "$use_miniconda" = "1" ] ; then wget https://repo.anaconda.com/miniconda/Miniconda3-py38_${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} pip; else \
    yum install -y python36 python36-devel python36-pip && \
    ln -sf /usr/bin/python36 /usr/bin/python && ln -s /usr/bin/pip3 /usr/bin/pip; \
    fi

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

RUN pip install -U pip==23.3.2

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

WORKDIR $HOME

# Arrow
# Need to enable c99 manually https://github.com/numpy/numpy/issues/14147
RUN bash -c 'if [ `arch` = "ppc64le" ]; then \
    ln -s /usr/bin/cmake3 /usr/bin/cmake && \
    git clone https://github.com/apache/arrow.git && \
    cd $HOME/arrow/cpp && \
    git checkout tags/apache-arrow-0.17.1 && \
    CFLAGS=-std=c99 pip install numpy==1.19.2 cython==0.29.14 scipy==1.5.2 && \
    export ARROW_BUILD_TYPE=release && \
    export ARROW_HOME=/usr/local && \
    cmake -DARROW_CXXFLAGS="-lutil" -DCMAKE_INSTALL_PREFIX=$ARROW_HOME -DARROW_PYTHON=on -DCMAKE_INSTALL_LIBDIR=lib -DCMAKE_BUILD_TYPE=$ARROW_BUILD_TYPE -DARROW_BUILD_TESTS=OFF && make -j && make install && \
    cd $HOME/arrow/python && \
    python setup.py install; \
    fi'

WORKDIR /

# Add requirements
COPY src/interface_py/requirements_runtime_demos_single_gpu.txt requirements_runtime_demos_single_gpu.txt
COPY src/interface_py/requirements_test.txt requirements_test.txt
RUN \
    chmod a+rwx requirements*.txt && \
    python -m pip install -r requirements_runtime_demos_single_gpu.txt && \
    python -m pip install -r requirements_test.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
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


# Add a canned jupyter notebook demo to the container
RUN \
    mkdir -p jupyter/demos
COPY examples/py/demos/H2O4GPU_Ridge.ipynb /jupyter/demos/H2O4GPU_Ridge.ipynb
COPY examples/py/demos/H2O4GPU_LinearRegression.ipynb /jupyter/demos/H2O4GPU_LinearRegression.ipynb
COPY examples/py/demos/H2O4GPU_GBM.ipynb /jupyter/demos/H2O4GPU_GBM.ipynb
COPY examples/py/demos/H2O4GPU_GLM.ipynb /jupyter/demos/H2O4GPU_GLM.ipynb
COPY examples/py/demos/H2O4GPU_Lasso.ipynb /jupyter/demos/H2O4GPU_Lasso.ipynb
COPY examples/py/demos/H2O4GPU_KMeans_Images.ipynb /jupyter/demos/H2O4GPU_KMeans_Images.ipynb
COPY examples/py/demos/H2O4GPU_KMeans_Quantization.ipynb /jupyter/demos/H2O4GPU_KMeans_Quantization.ipynb
COPY examples/py/demos/Multi-GPU-H2O-GLM-simple.ipynb /jupyter/demos/Multi-GPU-H2O-GLM-simple.ipynb
COPY examples/py/demos/H2O4GPU_TruncatedSVD.ipynb /jupyter/demos/H2O4GPU_TruncatedSVD.ipynb
COPY examples/py/demos/H2O4GPU_PCA.ipynb /jupyter/demos/H2O4GPU_PCA.ipynb
COPY examples/py/demos/H2O4GPU_Daal_LinearRegression.ipynb /jupyter/demos/H2O4GPU_Daal_LinearRegression.ipynb
COPY examples/py/demos/figures /jupyter/demos/figures
RUN \
    cd /jupyter/demos && \
    chmod -R a+rwx /jupyter && \
    mkdir /scikit_learn_data && \
    chmod -R a+rwx /scikit_learn_data && \
    mkdir -p /jupyter/scikit_learn_data/covertype && \
    chmod -R a+rwx /jupyter/scikit_learn_data/covertype && \
    mkdir -p /jupyter/scikit_learn_data/lfw_home && \
    chmod -R a+rwx /jupyter/scikit_learn_data/lfw_home && \
    HOME=/jupyter jupyter notebook --generate-config && \
    sed -i "s/#c.NotebookApp.token = '<generated>'/c.NotebookApp.token = 'h2o'/" /jupyter/.jupyter/jupyter_notebook_config.py && \
    chmod -R a+rwx /jupyter/.jupyter

# Add shell wrapper
COPY scripts/run.sh /run.sh
RUN \
    chmod a+rwx run.sh

ARG h2o4gpu_VERSION
ARG h2o4gpu_COMMIT
ARG DOCKER_VERSION_TAG
LABEL \
    h2o4gpu_commit="$h2o4gpu_COMMIT" \
    docker_version_tag="$DOCKER_VERSION_TAG"

ENTRYPOINT ["./run.sh"]
EXPOSE 8888