susynt-submit

Files and scripts needed to submit a SusyNt ntuple production

Instructions

Pre-requisites:

• a directory build_area where you have already built all the necessary packages. For this step, see susynt-write.
• localSetupROOT (with the same version used to compile build_area)
• voms-proxy-init and localSetupPandaClient

Quickstart

cd susynt-write/

setupATLAS
localSetupDQ2Client --skipConfirm
voms-proxy-init -voms atlas -valid 144:00
localSetupPandaClient currentJedi --noAthenaCheck
rcSetup
rc compile
cd ..

git clone git@github.com:gerbaudo/susynt-submit --branch xaod
cd susynt-submit
bash/create_tarball.sh

At this point you can submit the jobs for the samples you need. Below are some reference options:

./submit.py --filterTrig --contTau --nickname ${USER} -t ${TAG} -f txt/datasets/2015/data_egamma.txt
./submit.py --filterTrig --contTau --nickname ${USER} -t ${TAG} -f txt/datasets/2015/data_muons.txt
./submit.py --saveTruth  --contTau --nickname ${USER} -t ${TAG} -f txt/datasets/2015/mc_p1784.txt
# ./submit.py --saveTruth --filterOff --contTau --nickname ${USER} -t ${TAG} -f txt/signal/<list>.txt

Notes:

• always save container taus, --contTau
• for data it is useful to only write events that pass our triggers, --filterTrig; the reduction in simulated samples is probably not worth it
• for simulated samples, save the truth information, --saveTruth
• for signal samples, disable the lepton filter, --filterOff. This is useful when one needs to compute the signal efficiency, and the reduction would anyway be small.
• ${USER} and ${TAG} have to be specified. ${USER} must match your grid username.