The nextflow.config
file instructs Nextflow to use Apptainer with a custom Docker image, so the user only needs to run e.g.
./nextflow-sockeye.sh run funnel_scale.nf
for Sockeye or
./nextflow-alliance.sh run funnel_scale.nf
for the Alliance clusters.
If you have Julia and cmdstan installed locally, you can run
./nextflow run funnel_scale.nf
Note: this assumes that there is an environment variable CMDSTAN
pointing to the directory where cmdstan lives. You can do this via
export CMDSTAN=/full/path/to/cmdstan-vx.xx
Add the following at the bottom of your .bashrc
, replacing
YOUR_ALLOC_NAME
with the name of the allocation you are assigned to, andYOUR_GITHUB_KEY
with the filename of your private ssh key for accessing Github.
# export the name of the allocation you belong to
export ALLOC_NAME="YOUR_ALLOC_NAME"
# load modules
module load gcc/9.4.0
module load openjdk
module load git
module load apptainer
# start ssh-agent if not already started, then add keys
if [ -z "$SSH_AUTH_SOCK" ] ; then
eval `ssh-agent -s` > /dev/null
ssh-add $HOME/.ssh/YOUR_GITHUB_KEY >/dev/null 2>&1
fi