Repository navigation
Exclusive channel data preparation template
Andrey edited this page Feb 8, 2022
·
9 revisions
- output: HIPO format with all banks per event but only events with two pions are saved
- x6 factor data reduction compared to nSidis
- fall 2018 inbending: custom train (215 Gb) vs nSidis (1.2 Tb)
- frequency: once or twice
- production timeframe: 1 day maybe few, depending on farm
- 02/04/2022 run: 6 mins
- groovy script for (e,pi+,pi-,X) events: gist
- example of sbatch script
lvl1_epimpip.sbatch- script runs from the same directory where it was submitted
-
output files are created in the
epimpipdirectory which is expected to be there#!/bin/bash #SBATCH --ntasks=1 #SBATCH --job-name=twopionstrain #SBATCH --partition=production #SBATCH --account=clas12 #SBATCH --mem-per-cpu=1000 FILES=(/cache/clas12/rg-a/production/recon/fall2018/torus-1/pass1/v1/dst/train/nSidis/*) cd epimpip srun ../lvl1_epimpip.groovy ${FILES[$SLURM_ARRAY_TASK_ID]}
- submit 174 slurm jobs (one for each of the data files from directory in sbatch script):
sbatch -a 0-173 lvl1_epimpip.sbatch
- output: ROOT files with only relevant information
- data reduction:
- fall 2018 inbending: ROOT files (621 Mb) vs custom train (215 Gb) vs nSidis (1.2 Tb)
- frequency: occasionally at the beginning of the analysis, infrequently at the end
- production timeframe: few hours
- locations:
-
/work/clas12/kenjo/trains/root/eFPimFPiponly 1 electron (RGA cuts) only 1 Forward pi+ (RGA cuts) only 1 Forward pi- (RGA cuts) -
/work/clas12/kenjo/trains/root/eFPimFPipPonly 1 electron (RGA cuts) only 1 Forward pi+ (RGA cuts) only 1 Forward pi- (RGA cuts) only 1 proton (RGA cuts): FORWARD and CENTRAL -
/work/clas12/kenjo/trains/root/eFKpPimPonly 1 electron (RGA cuts) only 1 Forward K+ (RGA-ish cuts) only 1 pi- (RGA cuts): FORWARD and CENTRAL only 1 proton (RGA cuts): FORWARD and CENTRAL -
/work/clas12/kenjo/trains/root/eFKpKmPonly 1 electron (RGA cuts) only 1 Forward K+ (RGA-ish cuts) only 1 K- (RGA-ish cuts): FORWARD and CENTRAL only 1 proton (RGA cuts): FORWARD and CENTRAL
-
- Jupyter@jlab, laptop, desktop, ifarm
- time to process: minutes/hours