Simple Higgs Plot

Plots the mass distribution for H->ZZ*->4mu.

Installation

The program requires at least python 3.8 and ROOT 6.24 In addition an installed CMS Computing environment is required, in particular the Grid UI und CVMFS (Both available at CLIP).

Conda Bootstrap environment

As CLIP has no python3 installed, the recommended way is to use a bootstrap conda environment. The suggested environment contains some useful tools. In case you prefer, feel free to define your own.

Add to.bashrc

# Conda setup
eval "$('/software/2020/software/anaconda3/2019.10/bin/conda' 'shell.bash' 'hook' 2> /dev/null)"
conda activate /groups/hephy/cms/dietrich.liko/conda/envs/my-base

and login again.

Using CMSSW

As a relative modern toolchain is required, currently the release CMSSW_12_3_X_2022-02-02-1100 is suggested.

Create the environment

cmsrel CMSSW_12_3_X_2022-02-02-1100
cd CMS_12_3_X/src
cmsenv

and checkout the repository

git clone [email protected]:dietrichliko/higgs4l.git

Installation using Conda and Mamba

An alternative is to setup the full stack in a separate conda environment. It is suggested to move your private conda environments from the home area to the group area by creating a corresponding .condarc. Adapt the directory paths to your needs. It is suggested to move your private conda environments from the home area to the group area by creating a corresponding .condarc. Adapt the directory paths to your needs.

auto_activate_base: false
pip_interop_enabled: True
envs_dirs:
  - /groups/hephy/cms/dietrich.liko/conda/envs
pkgs_dirs:
  - /groups/hephy/cms/dietrich.liko/conda/pkgs

The bootstrap conda environment contains mamba, a faster implementation of conda install recommended for the use with conda-forge.

Create an environment for the Higgs analysis

conda activate my-base
mamba create -y -n higgs4l -c conda-forge python=3.8 root pyyaml
conda deactivate

Enable the environment enable it with conda

conda activate higgs4l

and checkout the repository

git clone [email protected]:dietrichliko/higgs4l.git

Running the program

Check the state of the CLIP scratch area. The full doubleMuon datasets requires about 850GB.

df -h /scratch/cbe

Pre-stage the data can be done from the login node

doublemuon_prestage.py

As the analysis program fills all CPU available, it is good practice to allocate a worker node. Please keep in mind that allocating a full node is relatively expensive and keep its usage to the required minimum.

srun --time=00:30:00 --nodes=1 --pty /bin/bash
./doublemuon.py