installOnHpc.md

The HPC cluster (University of Geneva) provides a wide range of software that can be linked using the module command. The official documentation is available on the UniGe website.

Environment Setup

The following chain of commands can be either set in the user .bash_profile or in a dedicated bash script.

echo "Loading ROOT module"
ml load GCC/11.2.0
ml load OpenMPI/4.1.1
module load ROOT

echo "Loading yaml-cpp module"
ml load GCCcore/11.3.0
module load yaml-cpp/0.7.0

echo "Loading other modules..."
module load nlohmann_json

module load Python
module load CMake
module load git
module load CUDA

The versions are automatically chosen to match the selected software.

Alright! You are now ready to compile :-).

Building Gundam

As for the general guide we propose to compile the software with this tree directory:

export WORK_DIR="$HOME"
export INSTALL_DIR="$WORK_DIR/install/"
export BUILD_DIR="$WORK_DIR/build/"
export REPO_DIR="$WORK_DIR/repositories/"

Then create these path:

mkdir -p $INSTALL_DIR
mkdir -p $BUILD_DIR
mkdir -p $REPO_DIR

Let's clone Gundam with git:

cd $REPO_DIR
git clone https://github.com/nadrino/gundam.git
cd $REPO_DIR/gundam
# fetch the latest tagged version
git checkout $(git describe --tags `git rev-list --tags --max-count=1`)
# don't forget to initialize the submodules
git submodule update --init --recursive

mkdir -p $BUILD_DIR/gundam
mkdir -p $INSTALL_DIR/gundam
cd $BUILD_DIR/gundam
cmake \
  -D CMAKE_INSTALL_PREFIX:PATH=$INSTALL_DIR/gundam \
  -D CMAKE_BUILD_TYPE=Release \
  $REPO_DIR/gundam/.
make install -j4

Once the software is installed, let's link the bin and lib folders:

export PATH="$INSTALL_DIR/gundam/bin:$PATH"
export LD_LIBRARY_PATH="$INSTALL_DIR/gundam/lib:$LD_LIBRARY_PATH"

You can put these lines in your bash profile as well.

Now you should be able to run your first gundam command:

gundamFitter

Congrats! Now gundam is installed on your cluster! :-D