search

ytatus94 / Higgsino

The NUHM2 Higgsino study
0 stars 0 forks source link

Joey's mail 20171114 #34

Open ytatus94 opened 6 years ago

ytatus94 commented 6 years ago

There are instructions for doing this here. I run most of my jobs on my batch system via Condor, with commands like:

run_xAODNtMaker --skipPRWCheck -selector HiggsinoSelector -groupSet SUSY16_Signal_RerunForFinalFits -d "/usatlas/groups/bnl_local/jreic/November1_2017_Higgsino_SUSY16_Signal_RerunForFinalFits_v1.9d_Skims/skims" -tag v1.9d --condor

run_makeDirectory -d "/usatlas/groups/bnl_local/jreic/November1_2017_Higgsino_SUSY16_Signal_RerunForFinalFits_v1.9d_Skims/" -tag v1.9d -groupSet SUSY16_Signal_RerunForFinalFits

moveCondorOutputsToGroupsetStructure -groupSet SUSY16_Signal_RerunForFinalFits -tag v1.9d -d /usatlas/groups/bnl_local/jreic/November1_2017_Higgsino_SUSY16_Signal_RerunForFinalFits_v1.9d_Skims/ -subdirPrefix skims

run_merge -groupSet SUSY16_Signal_RerunForFinalFits -d "/usatlas/groups/bnl_local/jreic/November1_2017_Higgsino_SUSY16_Signal_RerunForFinalFits_v1.9d_Skims/" -tag v1.9d --condor

where "SUSY16_Signal_RerunForFinalFits" is a groupSet that I only had locally; you could probably use SUSY16_Signal or make a dedicated SUSY16_Signal_NUHM2 which would be a copy of this directory but only containing the groupset.config and the relevant NUHM2 text files.

If you run via the grid instead of on a batch system, this "moveCondorOutputsToGroupsetStructure" step is unnecessary, and I believe is covered by run_download.

Our derivation skims on the MET trigger, so keep that in mind in trying to relax the MET cuts...

Joey

ytatus94 commented 6 years ago

Hi Yu-Ting,

This "LOOSE_NOMINAL" tree is a looser selection with no systematics, whereas the block you quoted is used for our NoSys ntuples as well as all systematic trees. But a few lines above is 

  // MET skimming 
  if(configMgr->obj->met.Et < 150.0) return false;
  configMgr->cutflow->bookCut("cutFlow","MET > 150 GeV", weight );

which you'll also want to comment out. In fact, you could probably just comment out this line and be done with it! (I believe this is what I did in the past).

Also, to only make NoSys trees rather than all of the systematics (which should be irrelevant for your study), you should go into your groupSet and change this: https://gitlab.cern.ch/atlas-phys-susy-higgsino/SusySkimHiggsino/blob/master/data/samples/SUSY16_Signal/groupset.config#L16 (but in whichever directory you actually are using!)

from:

SYSTEMATIC_SET

to

#SYSTEMATIC_SET

This tells the code to ignore systematics.

Joey