fnallpc-docker

This repository contains the Docker projects for the LHC Physics Center at Fermilab.

• Status
• Build Instructions
• Use Instructions
• Source Containers
• Conda Packages Included

Status

Branch	Build	Type	Pulls	Stars	Docker Hub
master
singularity-compatible					cms-cvmfs-docker

Build Instructions

Use Instructions

Source Containers

Conda Packages Included

Environment Name	Python 3.7 Container	Python 2.7 Container
Notes	--	--
arrow-cpp	0.15.1	0.11.1
awkward	0.12.20	0.12.20
bazel	0.26.1	0.26.1
blaze	0.11.3	0.11.3
bokeh	1.4.0	1.4.0
boost-cpp	1.71.0	--
cuda	9.0.176 *	9.0.176 *
cudatoolkit	10.1.243	10.1.243
cudnn	7.6.5	7.6.5
dask	2.10.1	1.2.2
EnergyFlow	1.0.2	1.0.2
graphviz	2.40.1	2.40.1
h5py	2.10.0	2.9.0
intel-openmp	2020.0	2020.0
ipython	7.12.0	5.8.0
jupyter_client	5.3.4	5.3.4
jupyter_core	4.6.1	4.6.1
keras-applications	1.0.8	1.0.8
keras-base	2.2.4	2.2.4
keras-gpu	2.2.4	2.2.4
keras-preprocessing	1.1.0	1.1.0
libboost	1.71.0	1.67.0
matplotlib	3.1.3	2.2.3
mkl	2020.0	2020.0
mplhep	0.0.35	0.0.35
nb_conda	2.2.1	2.2.1
numba	0.48.0	0.47.0
numpy	1.18.1	1.16.6
pandas	1.0.0	0.24.2
pip	20.0.2	19.3.1
pyarrow	0.15.1	0.11.1
pycuda	2019.1.2	2019.1.2
pydot	1.4.1	1.4.1
pygpu	0.7.6	0.7.6
python	3.7.6	2.7.17
PyTorch	1.3.1	1.3.1
root	--	--
rootpy	--	--
scikit-image	0.16.2	0.14.2
scikit-learn	0.22.1	0.20.3
scipy	1.4.1	1.2.1
tensorboard	1.14.0	1.14.0
tensorflow	1.14.0	1.14.0
tensorflow-base	1.14.0	1.14.0
tensorflow-estimator	1.14.0	1.14.0
tensorflow-gpu	1.14.0	1.14.0
Theano	1.0.4	1.0.4
torchvision	0.4.2	0.4.2
uproot	3.11.2	3.11.2
uproot-methods	0.7.3	0.7.3
wheel	0.34.2	0.33.6
yaml	0.1.7	0.1.7

To Run Using Docker

docker run --rm -it -P --gpus all --device /dev/fuse --net=host -e DISPLAY=host.docker.internal:0 -e MY_UID=501 -e MY_GID=20 tensorflow/tensorflow:latest-gpu-py3-jupyter docker run --rm -it -P --gpus all --device /dev/fuse --net=host -e DISPLAY=host.docker.internal:0 -e MY_UID=$(id -u) -e MY_GID=$(id -g) tensorflow/tensorflow:latest-gpu-py3

To Run Using Singularity

singularity shell --nv docker://fnallpc/fnallpc-docker:singularity-compatible #does not run bash for some reason
singularity exec --nv docker://fnallpc/fnallpc-docker:singularity-compatible /bin/bash
singularity run --nv docker://fnallpc/fnallpc-docker:singularity-compatible

To convert a SIF to a sandbox

singularity build --sandbox fnallpc-docker_singularity-compatible fnallpc-docker_singularity-compatible.sif

Singularity compatible branch

Add the following required lines to your login scripts:

# CUDA Setup
CUDA_VERSION_SHORT="${CUDA_VERSION%.*}"
export PATH=/usr/local/cuda-${CUDA_VERSION_SHORT}/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-${CUDA_VERSION_SHORT}/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

Optional: Add the following lines to your login scripts:

# add some aliases that don't exist by default in this container
alias lla='ll -ah'
alias voms-proxy-init='voms-proxy-init -voms cms --valid 192:00 -cert ~/.globus/usercert.pem -key ~/.globus/userkey.pem'

# Source some CMS/VOMS specific setup scripts
if [ -f "/cvmfs/cms.cern.ch/cmsset_default.sh" ]; then
    source /cvmfs/cms.cern.ch/cmsset_default.sh
else
    echo -e "Unable to source /cvmfs/cms.cern.ch/cmsset_default.sh"
fi
if [ -f "/cvmfs/oasis.opensciencegrid.org/mis/osg-wn-client/current/el6-x86_64/setup.sh" ]; then
    source /cvmfs/oasis.opensciencegrid.org/mis/osg-wn-client/current/el6-x86_64/setup.sh
else
    echo -e "Unable to setup the grid utilities from /cvmfs/oasis.opensciencegrid.org/"
fi

# Needed to access FNAL EOS
export XrdSecGSISRVNAMES="cmseos.fnal.gov"