Skip to content
Snippets Groups Projects
Commit 6fb40d9e authored by Patrick L.S. Connor's avatar Patrick L.S. Connor
Browse files

submitting 1M events for both FullSim and FastSim

parent 6a10d1c1
No related branches found
No related tags found
No related merge requests found
#!/bin/zsh #!/bin/zsh
nevents=$1 nevents=1000
id=$2 id=$1
cd $id cd $id
...@@ -13,10 +13,15 @@ py=${cfi}_GEN_SIM.py ...@@ -13,10 +13,15 @@ py=${cfi}_GEN_SIM.py
ls $py ls $py
echo "process.RandomNumberGeneratorService.generator.initialSeed = cms.untracked.uint32($id)" >> $py echo "process.RandomNumberGeneratorService.generator.initialSeed = cms.untracked.uint32($((id+1)))" >> $py
cmsRun $py cmsRun $py
cmsDriver.py step2 --conditions auto:run2_mc -s DIGI:pdigi_valid,L1,DIGI2RAW,HLT:@relval2016 --datatier GEN-SIM-DIGI-RAW-HLTDEBUG -n $nevents --era Run2_2016 --eventcontent FEVTDEBUGHLT --filein file:step1.root --fileout file:step2.root #> step2_TTbar_13+TTbar_13+DIGIUP15+RECOUP15+HARVESTUP15+ALCATTUP15.log 2>&1 cmsDriver.py step2 --conditions auto:run2_mc -s DIGI:pdigi_valid,L1,DIGI2RAW,HLT:@relval2016 --datatier GEN-SIM-DIGI-RAW-HLTDEBUG -n $nevents --era Run2_2016 --eventcontent FEVTDEBUGHLT --filein file:step1.root --fileout file:step2.root #> step2_TTbar_13+TTbar_13+DIGIUP15+RECOUP15+HARVESTUP15+ALCATTUP15.log 2>&1
cmsDriver.py step2 --conditions auto:run2_mc -s DIGI:pdigi_valid,L1,DIGI2RAW,HLT:@relval2016 --datatier GEN-SIM-DIGI-RAW-HLTDEBUG -n 1000 --era Run2_2016 --eventcontent FEVTDEBUGHLT --filein file:step1.root --fileout file:step2.root > step2_TTbar_13+TTbar_13+DIGIUP15+RECOUP15+HARVESTUP15+ALCATTUP15.log 2>&1
cmsDriver.py step3 --runUnscheduled --conditions auto:run2_mc -s RAW2DIGI,L1Reco,RECO,RECOSIM,EI,PAT,VALIDATION:@standardValidation+@miniAODValidation,DQM:@standardDQM+@ExtraHLT+@miniAODDQM --datatier GEN-SIM-RECO,AODSIM,MINIAODSIM,DQMIO -n $nevents --era Run2_2016 --eventcontent RECOSIM,AODSIM,MINIAODSIM,DQM --filein file:step2.root --fileout file:step3.root #> step3_TTbar_13+TTbar_13+DIGIUP15+RECOUP15+HARVESTUP15+ALCATTUP15.log 2>&1 cmsDriver.py step3 --runUnscheduled --conditions auto:run2_mc -s RAW2DIGI,L1Reco,RECO,RECOSIM,EI,PAT,VALIDATION:@standardValidation+@miniAODValidation,DQM:@standardDQM+@ExtraHLT+@miniAODDQM --datatier GEN-SIM-RECO,AODSIM,MINIAODSIM,DQMIO -n $nevents --era Run2_2016 --eventcontent RECOSIM,AODSIM,MINIAODSIM,DQM --filein file:step2.root --fileout file:step3.root #> step3_TTbar_13+TTbar_13+DIGIUP15+RECOUP15+HARVESTUP15+ALCATTUP15.log 2>&1
...@@ -17,9 +17,17 @@ cmsrel CMSSW_10_6_22 ...@@ -17,9 +17,17 @@ cmsrel CMSSW_10_6_22
Finally clone this repository in `$CMSSW_BASE/src`. Finally clone this repository in `$CMSSW_BASE/src`.
### In case of change of CMSSW version
The version of CMSSW may change in the future.
To track down the releases for the local architecture, just enter:
```
ls -1 -d /cvmfs/cms.cern.ch/*/cms/cmssw/CMSSW_10_6_22/src
```
## Execution ## Execution
Load the environement (necessary each time you open a new shell): Start a new shell (to avoid possible conflicts), and load the environement:
``` ```
source init source init
``` ```
...@@ -46,7 +54,8 @@ Just do: ...@@ -46,7 +54,8 @@ Just do:
./parallel ./parallel
``` ```
No option is necessary. No option is necessary.
You can change the number of events You can change the number of events in the script itself.
Beware that each time you run this command, the former root files are removed.
This approach can be useful to ensure that different seed are used for each job (the second option of `run`, which we previously ignored). This approach can be useful to ensure that different seed are used for each job (the second option of `run`, which we previously ignored).
Always check the occupancy of the local machine with `htop`, and don't go for this option if too many people are on this machine. Always check the occupancy of the local machine with `htop`, and don't go for this option if too many people are on this machine.
...@@ -58,6 +67,9 @@ Similar, just another script: ...@@ -58,6 +67,9 @@ Similar, just another script:
./submit ./submit
``` ```
This should be the privileged approach for large-scale production. This should be the privileged approach for large-scale production.
Here too, remember that each time you rerun the command, you actually remove the former run.
In case you want to extend the statistics of some existing sample, just clone this repo and run it from scratch.
#### Troubleshooting #### Troubleshooting
...@@ -70,8 +82,16 @@ If your job is on hold and you want to know more: ...@@ -70,8 +82,16 @@ If your job is on hold and you want to know more:
``` ```
condor_q -global -better-analyze JOBID condor_q -global -better-analyze JOBID
``` ```
(You get the job id when running `condor_q`.)
If a or several jobs was or were put on hold, and if you could fix the issue, then you can release the job as follows:
```
condor_release -all
```
If you want to kill all your jobs: If you want to kill all your jobs:
``` ```
condor_rm -all condor_rm -all
``` ```
Otherwise, consult the [official documentation](https://htcondor.readthedocs.io/en/latest/man-pages/index.html).
...@@ -4,8 +4,9 @@ export NJOBS=2 ...@@ -4,8 +4,9 @@ export NJOBS=2
for i in {1..$NJOBS} for i in {1..$NJOBS}
do do
rm -rf $i j=$((i-1))
mkdir $i rm -rf $j
./run $i & mkdir $j
./run $j &
done done
wait wait
#!/bin/zsh #!/bin/zsh
eval `/usr/bin/modulecmd zsh use -a /afs/desy.de/group/cms/modulefiles/` #eval `/usr/bin/modulecmd zsh use -a /afs/desy.de/group/cms/modulefiles/`
eval `/usr/bin/modulecmd zsh load cmssw` #eval `/usr/bin/modulecmd zsh load cmssw`
eval `scramv1 runtime -sh` #eval `scramv1 runtime -sh`
export LD_LIBRARY_PATH_STORED=$LD_LIBRARY_PATH export LD_LIBRARY_PATH_STORED=$LD_LIBRARY_PATH
export NJOBS=1000 export NJOBS=1000
for i in {1..$NJOBS} for i in {1..$NJOBS}
do do
rm -rf $i j=$((i-1))
mkdir $i rm -rf $j
mkdir $j
done done
condor_submit job condor_submit -batch-name ${PWD##*/} job
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please to comment