Dockerfile

################################################################################
# Notes: this image is large and many improvements are possible. 
# Sources:
# - https://uwekorn.com/2021/03/01/deploying-conda-environments-in-docker-how-to-do-it-right.html
# - https://pythonspeed.com/articles/conda-docker-image-size/
# micromamba is failing for PySR, so sticking with mambaforge for now.
# FROM --platform=linux/amd64 mambaorg/micromamba:0.21.2 as build
FROM condaforge/mambaforge:4.11.0-2 as base
################################################################################
# Nvidia code ##################################################################
################################################################################
ENV PATH /usr/local/nvidia/bin/:$PATH
ENV LD_LIBRARY_PATH /usr/local/nvidia/lib:/usr/local/nvidia/lib64:$LD_LIBRARY_PATH
# Tell nvidia-docker the driver spec that we need as well as to
# use all available devices, which are mounted at /usr/local/nvidia.
# The LABEL supports an older version of nvidia-docker, the env
# variables a newer one.
ENV NVIDIA_VISIBLE_DEVICES all
ENV NVIDIA_DRIVER_CAPABILITIES compute,utility
LABEL com.nvidia.volumes.needed="nvidia_driver"
################################################################################

# Install base packages.
USER root

ARG DEBIAN_FRONTEND=noninteractive

RUN apt update && apt install -y \
    default-jdk \
    rsync \
    # bzip2 \
    # ca-certificates \
    curl \
    # git \
    # wget \
    vim \
    jq && \
    rm -rf /var/lib/apt/lists/*

# Install env
################################################################################
USER $MAMBA_USER
SHELL ["/bin/bash", "-c"]
COPY --chown=$MAMBA_USER:$MAMBA_USER environment.yml /tmp/environment.yml
RUN --mount=type=cache,target=/opt/conda/pkgs mamba env create -f /tmp/environment.yml 
SHELL ["mamba", "run", "-n", "srbench", "/bin/bash", "-c"]
COPY . .
RUN bash install.sh