L1TSignalZerobiasMixer

This package contains instructions for

1. producing Madgraph5 based signals,
2. hadronize them using Pythia8,
3. simulate for CMS detector response using Delphes,
4. read zerobias L1TCaloRegion data, in a format suitable for L1TAutoEncoder studies, from separtately produced files in CMSSW runs,
5. overlay a random zerobias event with the produced signal,
6. save the data back in a format suitable for L1TAutoEncoder studies.

The inputs are madgraph, pythia8 and Delphes control files for the signal root file producer are named *.txt.

For the region dumper, the inputs are the root file from Delphes and the Zerobias file from the L1TRegionDumper.

Code installation instructions:

First time only

If you are on machines with /cvmfs and CentOS7 (login.hep.wisc.edu), you may use ROOT from there:

source /cvmfs/sft.cern.ch/lcg/views/LCG_101/x86_64-centos7-gcc11-opt/setup.sh 

If you are on machines with /cvmfs and CentOS8 (mucol01.hep.wisc.edu), you may use ROOT from there:

source /cvmfs/sft.cern.ch/lcg/views/LCG_101/x86_64-centos8-gcc11-opt/setup.sh

If you are on a Apple MacOS system, you need gcc, gfortran and root installed on your system, and you may need this additionally for Madgraph and Delphes work:

if [ `uname` == 'Darwin' ]; then echo export MACOSX_DEPLOYMENT_TARGET=10.15; fi

If you do not have a working Madgraph5 installation, do the following, in a directory with plenty of space:

wget https://launchpad.net/mg5amcnlo/3.0/3.2.x/+download/MG5_aMC_v3.2.0.tar.gz
tar zxf MG5_aMC_v3.2.0.tar.gz 
export mg5dir=$PWD/MG5_aMC_v3_2_0/
python $mg5dir/bin/mg5_aMC

From within the MG5_aMC prompt execute the following. They take a long time 10 mins to finish. Optionally, you can track the log files in the secondary login window, if you wish.

install lhapdf6
install pythia8
install Delphes
exit

If you do have a Madgraph5 directory already setup:

export mg5dir=<your MG5_aMC_v3_2_0 directory>

Go to the dirctory where you wish to work and then install this code:

if [ -d '/nfs_scratch' ]; then export basedir=/nfs_scratch/$USER/`date +%Y-%m-%d`; else basedir=$PWD/`date +%Y-%m-%d`; fi
mkdir -p $basedir
git clone git@github.com:SridharaDasu/L1TSignalZerobiasMixer.git
source $basedir/L1TSignalZerobiasMixer/setup.sh

On relogin use the base directory, i.e., the directory with the date of creation above, e.g., /nfs_scratch/dasu/2021-11-03/

source /nfs_scratch/dasu/2021-11-03/L1TSignalZerobiasMixer/setup.sh

To produce signal data (root files) use *.txt files with different configurations; If you make your own signal process files, please share by making a pull request

cd $datadir
python $mg5dir/bin/mg5_aMC $workdir/cms-vbfh-pythia8-delphes.txt

With above .txt file the event files will be in the directory $datadir/cms-vbfh-pythia8-delphes/run_01/

There should now be a root file in your directory $datadir/cms-vbfh-pythia8-delphes/Events/run_01/tag_1_delphes_events.root

We use root to read this data, and the CSV file from the zerobias run (https://github.com/SridharaDasu/L1TRegionDumper) to produce the final file:

export SIGNAL_ROOT_FILE=$datadir/cms-vbfh-pythia8-delphes/Events/run_01/tag_1_delphes_events.root
export ZEROBIAS_CSV_FILE=/nfs_scratch/dasu/2022-01/CMSSW_11_1_9/src/L1Trigger/L1TRegionDumper/test/L1TRegionDump.csv
export OUTPUT_CSV_FILE="$datadir/cms-vbfh.csv"
cd $workdir
root -l -q L1TSignalZerobiasMixer.C\(\"$SIGNAL_ROOT_FILE\"\,\"$ZEROBIAS_CSV_FILE\"\,\"$OUTPUT_CSV_FILE\"\)

Command to run Madgraph on the UW cluster:

runWiscJobs.py \
  --WorkFlow MG5Jobs \
  --Executable=runMG5JobOnWorker.sh \
  --Arguments=cms-vbfh-pythia8-delphes.txt \
  --nJobs=10 \
  --TransferInputFile=/nfs_scratch/dasu/CentOS7/MyMG5Dir.tar.gz,cms-vbfh-pythia8-delphes.txt \
  --OutputDir=/nfs_scratch/$USER \
  --HDFSProdDir None \
  --Experiment mucol \
  --MemoryRequirements 2048