MG5_Pythia8_Delphes_tutorial

This tutorial contains a specific guide to

1. Generate hard events using MadGraph5
2. Hadronize using Pythia8
3. Run fast simulation for FCC-hh (will be compatible for other collider) detector's response using Delphes
4. Submit jobs with Slurm in SLAC S3DF computing environment to produce large number of events

Installation (first time only)

I highly recommend to run this tutorial and all Monte_Carlo generation at S3DF iana /fs/ddn/sdf/group/atlas/d/$USER for large storage space.

First, clone this repo to hold everything

cd /fs/ddn/sdf/group/atlas/d/$USER
git clone https://github.com/Kenny-Jia/MG5_Pythia8_Delphes_tutorial Monte_Carlo
export workdir=$PWD
cd $workdir

If you are on either S3DF or lxplus, you should have access to cvmfs. We will use ROOT from there.

source /cvmfs/sft.cern.ch/lcg/views/LCG_105/x86_64-centos7-gcc11-opt/setup.sh

Download the Madgragh5 package from the official website

wget https://launchpad.net/mg5amcnlo/3.0/3.5.x/+download/MG5_aMC_v3.5.6.tar.gz
tar zxf MG5_aMC_v3.5.6.tar.gz 
export mg5dir=$PWD/MG5_aMC_v3_5_6/

Open mg5_aMC

python $mg5dir/bin/mg5_aMC

Now let's install all kinds of package we need within MG5_aMC prompt. Be patient, they may take like 25 mins to finish.

install lhapdf6
install pythia8
install Delphes
exit

So far, this is only design for generating FCC-hh sample with pileup. But you can easily switch from the default to any detector card like what we need to do for FCC-hh here:

cp $mg5dir/Delphes/cards/FCC/FCChh_PileUp.tcl $mg5dir/Template/Common/Cards/delphes_card_default.dat

Running with local machine

Assume you have successfully install everything, here the example is generating FCChh diHiggs production as an example.

source setup.sh
cd $datadir
python $mg5dir/bin/mg5_aMC $workdir/MG_script/example-MG5-pythia8-delphes.txt 

Running with cluster using slurm

Similarly, we use the example MG5 script card:

cd $workdir/Monte_Carlo
source setup.sh
sbatch $workdir/Monte_Carlo/submit_job.sh

You can check the progress with

squeue -u $USER

Optional, Debugging, and other issues

pip install readline

readline is optional but strongly recommended for interactive MadGraph usage. For any issues regarding this repo, please feel free to contact Kenny through email: [email protected]. I would be happy to help!