Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

V1.7 cuda 12.3 ubuntu 22.04 #144

Merged
merged 4 commits into from
Apr 9, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
132 changes: 69 additions & 63 deletions .build/Dockerfile
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@

# Use NVIDIA CUDA as base image and run the same installation as in the other packages.
# The version of cuda must match those of the packages installed in src/Dockerfile.gpulibs
FROM nvidia/cuda:12.0.1-cudnn8-runtime-ubuntu22.04
FROM nvidia/cuda:12.3.2-cudnn9-runtime-ubuntu22.04
LABEL authors="Christoph Schranz <christoph.schranz@salzburgresearch.at>, Mathematical Michael <consistentbayes@gmail.com>"
# This is a concatenated Dockerfile, the maintainers of subsequent sections may vary.
RUN chmod 1777 /tmp && chmod 1777 /var/tmp
Expand Down Expand Up @@ -43,22 +43,23 @@ USER root
# but lacks all features (e.g., download as all possible file formats)
ENV DEBIAN_FRONTEND noninteractive
RUN apt-get update --yes && \
# - `apt-get upgrade` is run to patch known vulnerabilities in apt-get packages as
# the Ubuntu base image is rebuilt too seldom sometimes (less than once a month)
# - `apt-get upgrade` is run to patch known vulnerabilities in system packages
# as the Ubuntu base image is rebuilt too seldom sometimes (less than once a month)
apt-get upgrade --yes && \
apt-get install --yes --no-install-recommends \
# - bzip2 is necessary to extract the micromamba executable.
bzip2 \
ca-certificates \
locales \
sudo \
# - tini is installed as a helpful container entrypoint that reaps zombie
# processes and such of the actual executable we want to start, see
# https://github.com/krallin/tini#why-tini for details.
# - `tini` is installed as a helpful container entrypoint,
# that reaps zombie processes and such of the actual executable we want to start
# See https://github.com/krallin/tini#why-tini for details
tini \
wget && \
apt-get clean && rm -rf /var/lib/apt/lists/* && \
echo "en_US.UTF-8 UTF-8" > /etc/locale.gen && \
echo "C.UTF-8 UTF-8" >> /etc/locale.gen && \
locale-gen

# Configure environment
Expand All @@ -67,9 +68,9 @@ ENV CONDA_DIR=/opt/conda \
NB_USER="${NB_USER}" \
NB_UID=${NB_UID} \
NB_GID=${NB_GID} \
LC_ALL=en_US.UTF-8 \
LANG=en_US.UTF-8 \
LANGUAGE=en_US.UTF-8
LC_ALL=C.UTF-8 \
LANG=C.UTF-8 \
LANGUAGE=C.UTF-8
ENV PATH="${CONDA_DIR}/bin:${PATH}" \
HOME="/home/${NB_USER}"

Expand Down Expand Up @@ -123,34 +124,33 @@ RUN set -x && \
# Should be simpler, see <https://github.com/mamba-org/mamba/issues/1437>
arch="64"; \
fi && \
wget --progress=dot:giga -O /tmp/micromamba.tar.bz2 \
"https://micromamba.snakepit.net/api/micromamba/linux-${arch}/latest" && \
tar -xvjf /tmp/micromamba.tar.bz2 --strip-components=1 bin/micromamba && \
rm /tmp/micromamba.tar.bz2 && \
# https://mamba.readthedocs.io/en/latest/installation/micromamba-installation.html#linux-and-macos
wget --progress=dot:giga -O - \
"https://micro.mamba.pm/api/micromamba/linux-${arch}/latest" | tar -xvj bin/micromamba && \
PYTHON_SPECIFIER="python=${PYTHON_VERSION}" && \
if [[ "${PYTHON_VERSION}" == "default" ]]; then PYTHON_SPECIFIER="python"; fi && \
# Install the packages
./micromamba install \
./bin/micromamba install \
--root-prefix="${CONDA_DIR}" \
--prefix="${CONDA_DIR}" \
--yes \
"${PYTHON_SPECIFIER}" \
'mamba' \
'jupyter_core' && \
rm micromamba && \
rm -rf /tmp/bin/ && \
# Pin major.minor version of python
mamba list python | grep '^python ' | tr -s ' ' | cut -d ' ' -f 1,2 >> "${CONDA_DIR}/conda-meta/pinned" && \
# https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-pkgs.html#preventing-packages-from-updating-pinning
mamba list --full-name 'python' | tail -1 | tr -s ' ' | cut -d ' ' -f 1,2 | sed 's/\.[^.]*$/.*/' >> "${CONDA_DIR}/conda-meta/pinned" && \
mamba clean --all -f -y && \
fix-permissions "${CONDA_DIR}" && \
fix-permissions "/home/${NB_USER}"

# Configure container startup
ENTRYPOINT ["tini", "-g", "--"]
CMD ["start.sh"]

# Copy local files as late as possible to avoid cache busting
COPY run-hooks.sh start.sh /usr/local/bin/

# Configure container entrypoint
ENTRYPOINT ["tini", "-g", "--", "start.sh"]

USER root

# Create dirs for startup hooks
Expand Down Expand Up @@ -181,16 +181,18 @@ SHELL ["/bin/bash", "-o", "pipefail", "-c"]

USER root

# Install all OS dependencies for the Server that starts but lacks all
# features (e.g., download as all possible file formats)
# Install all OS dependencies for the Server that starts
# but lacks all features (e.g., download as all possible file formats)
RUN apt-get update --yes && \
apt-get install --yes --no-install-recommends \
# - Add necessary fonts for matplotlib/seaborn
# See https://github.com/jupyter/docker-stacks/pull/380 for details
fonts-liberation \
# - pandoc is used to convert notebooks to html files
# - `pandoc` is used to convert notebooks to html files
# it's not present in the aarch64 Ubuntu image, so we install it here
pandoc \
# - run-one - a wrapper script that runs no more
# than one unique instance of some command with a unique set of arguments,
# - `run-one` - a wrapper script that runs no more
# than one unique instance of some command with a unique set of arguments,
# we use `run-one-constantly` to support the `RESTARTABLE` option
run-one && \
apt-get clean && rm -rf /var/lib/apt/lists/*
Expand Down Expand Up @@ -234,7 +236,7 @@ RUN fix-permissions /etc/jupyter/
# HEALTHCHECK documentation: https://docs.docker.com/engine/reference/builder/#healthcheck
# This healtcheck works well for `lab`, `notebook`, `nbclassic`, `server`, and `retro` jupyter commands
# https://github.com/jupyter/docker-stacks/issues/915#issuecomment-1068528799
HEALTHCHECK --interval=5s --timeout=3s --start-period=5s --retries=3 \
HEALTHCHECK --interval=3s --timeout=1s --start-period=3s --retries=3 \
CMD /etc/jupyter/docker_healthcheck.py || exit 1

# Switch back to jovyan to avoid accidental container runs as root
Expand Down Expand Up @@ -271,10 +273,10 @@ RUN apt-get update --yes && \
vim-tiny \
# git-over-ssh
openssh-client \
# less is needed to run help in R
# `less` is needed to run help in R
# see: https://github.com/jupyter/docker-stacks/issues/1588
less \
# nbconvert dependencies
# `nbconvert` dependencies
# https://nbconvert.readthedocs.io/en/latest/install.html#installing-tex
texlive-xetex \
texlive-fonts-recommended \
Expand Down Expand Up @@ -383,36 +385,36 @@ WORKDIR "${HOME}"

LABEL maintainer="Christoph Schranz <christoph.schranz@salzburgresearch.at>, Mathematical Michael <consistentbayes@gmail.com>"

# Install dependencies for e.g. PyTorch
RUN mamba install --quiet --yes \
pyyaml setuptools cmake cffi typing && \
mamba clean --all -f -y && \
fix-permissions "${CONDA_DIR}" && \
fix-permissions "/home/${NB_USER}"

# Install Tensorflow, check compatibility here:
# https://www.tensorflow.org/install/source#gpu
# installation via conda leads to errors in version 4.8.2
# Install CUDA-specific nvidia libraries and update libcudnn8 before that
# using device_lib.list_local_devices() the cudNN version is shown, adapt version to tested compat
USER ${NB_UID}
RUN pip install --upgrade pip && \
pip install --no-cache-dir tensorflow==2.15.0 keras==2.15.0 && \
fix-permissions "${CONDA_DIR}" && \
fix-permissions "/home/${NB_USER}"

# Install PyTorch with dependencies
RUN mamba install --quiet --yes \
pyyaml setuptools cmake cffi typing && \
mamba clean --all -f -y && \
pip install --no-cache-dir tensorflow==2.16.1 keras==3.1.1 && \
fix-permissions "${CONDA_DIR}" && \
fix-permissions "/home/${NB_USER}"

# Check compatibility here:
# https://pytorch.org/get-started/locally/
# Installation via conda leads to errors installing cudatoolkit=11.1
# RUN pip install --no-cache-dir torch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 && \
# torchviz==0.0.2 --extra-index-url https://download.pytorch.org/whl/cu118
# RUN pip install --no-cache-dir torch==2.2.2 torchvision==0.17.2 torchaudio==2.2.2 \
# && torchviz==0.0.2 --extra-index-url https://download.pytorch.org/whl/cu121
RUN set -ex \
&& buildDeps=' \
torch==2.1.2 \
torchvision==0.16.2 \
torchaudio==2.1.2 \
torch==2.2.2 \
torchvision==0.17.2 \
torchaudio==2.2.2 \
' \
&& pip install --no-cache-dir $buildDeps \
&& pip install --no-cache-dir $buildDeps --extra-index-url https://download.pytorch.org/whl/cu121 \
&& fix-permissions "${CONDA_DIR}" \
&& fix-permissions "/home/${NB_USER}"

Expand Down Expand Up @@ -441,7 +443,8 @@ RUN pip install --no-cache-dir nvidia-pyindex && \
fix-permissions "${CONDA_DIR}" && \
fix-permissions "/home/${NB_USER}"

RUN mamba install -c nvidia cuda-nvcc=12.0.140 -y && \
# Install cuda-nvcc with sepecific version, see here: https://anaconda.org/nvidia/cuda-nvcc/labels
RUN mamba install -c nvidia cuda-nvcc=12.3.107 -y && \
mamba clean --all -f -y && \
fix-permissions $CONDA_DIR && \
fix-permissions /home/$NB_USER
Expand All @@ -467,39 +470,42 @@ RUN apt-get update \
USER $NB_UID
RUN set -ex \
&& buildDeps=' \
graphviz==0.19.1 \
pytest==7.2.2 \
' \
graphviz==0.20.3 \
pytest==8.1.1 \
' \
&& pip install --no-cache-dir $buildDeps \
&& fix-permissions "${CONDA_DIR}" \
&& fix-permissions "/home/${NB_USER}"

# upgrade jupyter-server for compatibility
RUN pip install --no-cache-dir --upgrade \
distributed==2023.3.0 \
jupyter-server==2.4 \
# fix permissions of conda
RUN set -ex \
&& buildDeps=' \
distributed==2024.4.1 \
jupyter-server==2.13 \
' \
&& pip install --no-cache-dir $buildDeps \
&& fix-permissions "${CONDA_DIR}" \
&& fix-permissions "/home/${NB_USER}"

RUN pip install --no-cache-dir \
RUN set -ex \
&& buildDeps=' \
# install extension manager
jupyter_contrib_nbextensions==0.7.0 \
jupyter_nbextensions_configurator==0.6.1 \
jupyter_nbextensions_configurator==0.6.3 \
# install git extension
jupyterlab-git==0.41.0 \
jupyterlab-git==0.50.0 \
# install plotly extension
plotly==5.13.1 \
# install drawio and graphical extensions
jupyterlab-drawio==0.9.0 \
rise==5.7.1 \
ipyleaflet==0.17.2 \
ipywidgets==8.0.4 \
plotly==5.20.0 \
# install drawio and graphical extensions, not compatible with Jupyterlab 4.X yet
# ipydrawio==1.3.0 \
ipyleaflet==0.18.2 \
ipywidgets==8.1.2 \
# install spell checker
jupyterlab-spellchecker==0.7.3 && \
# fix permissions of conda
fix-permissions "${CONDA_DIR}" && \
fix-permissions "/home/${NB_USER}"
jupyterlab-spellchecker==0.8.4 \
' \
&& pip install --no-cache-dir $buildDeps \
&& fix-permissions "${CONDA_DIR}" \
&& fix-permissions "/home/${NB_USER}"

# Switch back to jovyan to avoid accidental container runs as root
USER $NB_UID
Expand Down
2 changes: 1 addition & 1 deletion .build/docker-stacks
Submodule docker-stacks updated 67 files
+0 −278 .github/actions/download-manifests/action.yml
+6 −3 .github/actions/load-image/action.yml
+0 −4 .github/dependabot.yml
+1 −1 .github/workflows/contributed-recipes.yml
+40 −16 .github/workflows/docker-build-test-upload.yml
+11 −6 .github/workflows/docker-merge-tags.yml
+22 −4 .github/workflows/docker-tag-push.yml
+10 −10 .github/workflows/docker-wiki-update.yml
+123 −60 .github/workflows/docker.yml
+13 −11 .pre-commit-config.yaml
+1 −1 .readthedocs.yaml
+15 −15 Makefile
+13 −11 README.md
+1 −1 aarch64-runner/setup.sh
+2 −2 binder/Dockerfile
+1 −1 docs/conf.py
+10 −9 docs/using/common.md
+1 −1 docs/using/recipe_code/oracledb.dockerfile
+11 −11 docs/using/recipes.md
+7 −7 docs/using/running.md
+13 −5 docs/using/selecting.md
+21 −14 docs/using/specifics.md
+1 −1 docs/using/troubleshooting.md
+8 −6 images/base-notebook/Dockerfile
+14 −1 images/base-notebook/docker_healthcheck.py
+10 −7 images/base-notebook/start-notebook.py
+6 −3 images/base-notebook/start-singleuser.py
+19 −19 images/docker-stacks-foundation/Dockerfile
+21 −2 images/docker-stacks-foundation/start.sh
+2 −2 images/minimal-notebook/Dockerfile
+6 −6 images/minimal-notebook/setup-scripts/setup-julia-packages.bash
+16 −4 images/minimal-notebook/setup-scripts/setup_julia.py
+20 −33 images/pyspark-notebook/Dockerfile
+2 −3 images/pyspark-notebook/ipython_kernel_config.py
+123 −0 images/pyspark-notebook/setup_spark.py
+1 −1 images/pytorch-notebook/Dockerfile
+25 −0 images/pytorch-notebook/cuda11/Dockerfile
+25 −0 images/pytorch-notebook/cuda12/Dockerfile
+3 −2 images/tensorflow-notebook/Dockerfile
+27 −0 images/tensorflow-notebook/cuda/Dockerfile
+9 −0 images/tensorflow-notebook/cuda/nvidia-lib-dirs.sh
+15 −2 tagging/apply_tags.py
+25 −0 tagging/get_prefix.py
+1 −2 tagging/images_hierarchy.py
+14 −4 tagging/merge_tags.py
+9 −19 tagging/taggers.py
+15 −4 tagging/update_wiki.py
+10 −3 tagging/write_manifest.py
+17 −4 tagging/write_tags_file.py
+6 −2 tests/R_mimetype_check.py
+1 −1 tests/all-spark-notebook/test_spark_notebooks.py
+5 −5 tests/base-notebook/test_container_options.py
+30 −30 tests/base-notebook/test_healthcheck.py
+1 −1 tests/base-notebook/test_pandoc.py
+2 −5 tests/base-notebook/test_start_container.py
+12 −9 tests/docker-stacks-foundation/test_packages.py
+12 −2 tests/docker-stacks-foundation/test_python_version.py
+1 −1 tests/docker-stacks-foundation/test_units.py
+32 −15 tests/docker-stacks-foundation/test_user_options.py
+10 −10 tests/minimal-notebook/data/notebook_math.ipynb
+3 −10 tests/minimal-notebook/test_nbconvert.py
+2 −2 tests/package_helper.py
+1 −1 tests/pluto_check.py
+1 −1 tests/run_command.py
+0 −1 tests/scipy-notebook/test_cython.py
+1 −1 tests/scipy-notebook/test_extensions.py
+1 −2 tests/scipy-notebook/test_matplotlib.py
15 changes: 14 additions & 1 deletion .build/docker_healthcheck.py
Original file line number Diff line number Diff line change
Expand Up @@ -3,14 +3,27 @@
# Distributed under the terms of the Modified BSD License.
import json
import os
import subprocess
from pathlib import Path

import requests

# Several operations below deliberately don't check for possible errors
# As this is a healthcheck, it should succeed or raise an exception on error

runtime_dir = Path("/home/") / os.environ["NB_USER"] / ".local/share/jupyter/runtime/"
# Docker runs healtchecks using an exec
# It uses the default user configured when running the image: root for the case of a custom NB_USER or jovyan for the case of the default image user.
# We manually change HOME to make `jupyter --runtime-dir` report a correct path
# More information: <https://github.com/jupyter/docker-stacks/pull/2074#issuecomment-1879778409>
result = subprocess.run(
["jupyter", "--runtime-dir"],
check=True,
capture_output=True,
text=True,
env=dict(os.environ) | {"HOME": "/home/" + os.environ["NB_USER"]},
)
runtime_dir = Path(result.stdout.rstrip())

json_file = next(runtime_dir.glob("*server-*.json"))

url = json.loads(json_file.read_bytes())["url"]
Expand Down
12 changes: 6 additions & 6 deletions .build/setup-scripts/setup-julia-packages.bash
Original file line number Diff line number Diff line change
Expand Up @@ -3,28 +3,28 @@ set -exuo pipefail
# Requirements:
# - Run as a non-root user
# - The JULIA_PKGDIR environment variable is set
# - Julia is already set up, with the setup-julia.bash command
# - Julia is already set up, with the setup_julia.py command


# If we don't specify what CPUs the precompilation should be done for, it's
# *only* done for the target of the host doing the compilation. When the
# container runs on a host that's the same architecture, but a *different*
# generation of CPU than what the build host was, the precompilation is useless
# and Julia takes a long long time to start up. This specific multitarget comes
# from https://github.com/JuliaCI/julia-buildkite/blob/70bde73f6cb17d4381b62236fc2d96b1c7acbba7/utilities/build_envs.sh#L20-L76,
# from https://github.com/JuliaCI/julia-buildkite/blob/9f354745a1f2bf31a5952462aa1ff2d869507cb8/utilities/build_envs.sh#L20-L82,
# and may need to be updated as new CPU generations come out.
# If the architecture the container runs on is different,
# precompilation may still have to be re-done on first startup - but this
# *should* catch most of the issues. See
# https://github.com/jupyter/docker-stacks/issues/2015 for more information
if [ "$(uname -m)" == "x86_64" ]; then
# See https://github.com/JuliaCI/julia-buildkite/blob/70bde73f6cb17d4381b62236fc2d96b1c7acbba7/utilities/build_envs.sh#L24
# See https://github.com/JuliaCI/julia-buildkite/blob/9f354745a1f2bf31a5952462aa1ff2d869507cb8/utilities/build_envs.sh#L23
# for an explanation of these options
export JULIA_CPU_TARGET="generic;sandybridge,-xsaveopt,clone_all;haswell,-rdrnd,base(1)"
export JULIA_CPU_TARGET="generic;sandybridge,-xsaveopt,clone_all;haswell,-rdrnd,base(1);x86-64-v4,-rdrnd,base(1)"
elif [ "$(uname -m)" == "aarch64" ]; then
# See https://github.com/JuliaCI/julia-buildkite/blob/70bde73f6cb17d4381b62236fc2d96b1c7acbba7/utilities/build_envs.sh#L54
# See https://github.com/JuliaCI/julia-buildkite/blob/9f354745a1f2bf31a5952462aa1ff2d869507cb8/utilities/build_envs.sh#L56
# for an explanation of these options
export JULIA_CPU_TARGET="generic;cortex-a57;thunderx2t99;carmel"
export JULIA_CPU_TARGET="generic;cortex-a57;thunderx2t99;carmel,clone_all;apple-m1,base(3);neoverse-512tvb,base(3)"
fi

# Install base Julia packages
Expand Down
Loading
Loading