Top Banner
GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007 . 5 0 * 10 4 ) 892 ( BR K B d 4 10 1 ) 1020 ( BR B s K K 0 * K K GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED 1K SUB-SAMPLES ON LXPLUS BUT LXPLUS TOO SLOW FOR LARGE SAMPLES – MOVE TO GRID? MC STUDY OF RADIATIVE B-DECAYS FOR B-PHYSICS/HLT STUDIES
17

GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

Dec 25, 2015

Download

Documents

Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

GETTING STARTED ON THE GRID:

W. G. SCOTT (RAL/PPD)RAL PHYSICS MEETINGTUES 15 MAY 2007

. 50* 104 )892( BRKBd

4101 )1020( BRBs

KK 0*

KK

GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006)SIMULATED/DIGITISDED 1K SUB-SAMPLES ON LXPLUSBUT LXPLUS TOO SLOW FOR LARGE SAMPLES – MOVE TO GRID?

MC STUDY OF RADIATIVE B-DECAYS FOR B-PHYSICS/HLT STUDIES

Page 2: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

csc.018200.B_Kstar_Kplus_PiMinus_gamma4.evgen._00001.pool.root csc.018200.B_kstar_Kplus_PiMinus_gamma4.evgen._00002.pool.root csc.018200.B_kstar_Kplus_PiMinus_gamma4.evgen._00003.pool.root csc.018200.B_kstar_Kplus_PiMinus_gamma4.evgen._00004.pool.root csc.018200.B_kstar_Kplus_PiMinus_gamma4.evgen._00005.pool.root csc.018200.B_Kstar_Kplus_PiMinus_gamma4.evgen._00006.pool.root csc.018200.B_Ksatr_Kplus_PiMinus_gamma4.evgen._00007.pool.root csc.018200.B_Ksatr_Kplus_PiMinus_gamma4.evgen._00008.pool.rootcsc.018200.B_kstar_Kplus_PiMinus_gamma4.evgen._00009.pool.rootcsc.018200.B_Ksatr_Kplus_piMinus_gamma4.evgen._00010.pool.root

/castor/cern.ch/user/s/scottw/gjab02/

10 files @ 1000 events per file = 10,000 B -> K* gamma events total

Generated Bd -> K* gamma Events stored in Castor:

Generation is easy: 1000 events per 8 hr job ~ 30Mb per file

RELEASE 12.0.0

csc no. is not present for these Bd -> K* gamma data sets

Page 3: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

*KBd GeV 4 )(

GeV 6 )(

T

T

P

P

GeV )(TP

Page 4: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

*KBd GeV 4 )(

GeV 6 )(

T

T

P

P

)( K

)(

K

Page 5: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00001.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00002.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00003.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00004.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00005.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00006.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00007.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00008.pool.rootcsc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00009.pool.rootcsc.018201.Bs_phi_Kplus_Kminus_gamma4.evgen._00010.pool.root

/castor/cern.ch/user/s/scottw/gjab02/

10 files @ 1000 events per file = 10,000 Bs -> Phi gamma events total

Bs -> phi gamma Events stored in Castor:

RELEASE 12.0.0

Generation is easy: 1000 events per 8 hr job ~ 30Mb per file

Page 6: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

sBGeV 4 )(

GeV 6 )(

T

T

P

P

GeV )(TP

Page 7: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

sBGeV 4 )(

GeV 6 )(

T

T

P

P

)( K

)(

K

Page 8: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

Simulated and digistsed Bd -> K* gamma Samples:

B_Kstar_Kplus_PiMinus_gamma4.digi._00001.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00002.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00003.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00004.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00005.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00006.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00007.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00008.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00009.pool.rootB_Kstar_Kplus_PiMinus_gamma4.digi._00010.pool.root

castor/cern.ch/user/s/scottw/gjab02/

10 files @ 100 events per file = 1000 Bd -> K* gamma events total

100 EVENTS PER 1 DAY JOB ON LXPLUS = TOO SLOW !!~200Mb PER 100 EVENT FILE!!

RELEASE 12.0.6

USING STANDARDSIMUL+DIGITTRANSFORMATIONS

Page 9: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

Simulated and digitised Bs -> phi gamma sample:

csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00001.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00002.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00003.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00004.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00005.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00006.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00007.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00008.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00009.pool.root csc.018201.Bs_phi_Kplus_Kminus_gamma4.digit._00010.pool.root

castor/cern.ch/user/s/scottw/gjab02/

10 files @ 100 events per file = 1000 Bs -> Phi gamma events total

100 EVENTS PER 1 DAY JOB ON LXPLUS = TOO SLOW !!~200Mb PER 100 EVENT FILE!!

RELEASE 12.0.6

USING STANDARDSIMUL+DIGITTRANSFORMATIONS

Page 10: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

LVL-! TRIGGER EFFIENCY PLOT (THANKS JULIE!)

LVL-1 THR:

3 GeV4 GeV

5 GeV

*KBd

GeV )(TP

Page 11: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

MORE JULIEPLOTS!!

Page 12: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

MOVING TO THE GRID:

/afs/cern.ch/user/s/scottw/mycertificate.pfx

You need a copy of your Grid Certificate in your home directory:

Go to your work directory and use z-shell:

cd public/b1206zsh

Execute the following commands to setup Ganga (and Don-Quijote):

source /afs/cern.ch/project/gd/LCG-share/sl3/etc/profile.d/grid_env.shsource /afs/usatlas.bnl.gov/Grid/Don-Quijote/dq2_user_client/setup.zsh.CERNexport PATH=${PATH}:/afs/cern.ch/sw/ganga/install/slc3_gcc323/4.2.12/bin/export GANGA_CONFIG_PATH=GangaAtlas/Atlas.inivoms-proxy-init

Enter your 18-character pass phrase in response to this last command:******************

ALL QUESTIONS TO S. DALLISON PLEASE!!

Page 13: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

PREPARE A SIMPLE JOB TO RUN ON THE GRID:

emacs mygangajob.py Prepare the “job” as a simple python script (in your work area)t:

j=Job()j.application=Executable()j.application.exe=File('$HOME/public/b1206/myscript.sh')j.backend=LCG()j.submit()

Prepare the “exe” referenced above as a simple shell script:emacs myscript.sh

#!/bin/shecho 'myscript.sh running...'echo "----------------------"/bin/hostnameecho "----------------------"env

This script does essentially nothingexcept to announce that it is runningand list all the environment variables

https://twiki.cern.ch/twiki/bin/view/Atlas/GangaTutorial427

Page 14: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

RUN THE JOB USING GANGA:

Start Ganga and submit the job:

ganga*** Welcome to Ganga ***Version: Ganga-4-2-12 ....etc.In[1]: execfile(`mygangajob.py’)submitting job 58 to LCG backend …. etc.In[2]:jobs (to see status of all jobs)In[3]:jobs[58] (to see detailed status of job 58)

job status changes:submitted /new?runningcompleted /failed

When the job is completed, vew the output using !less:

In[4]!less /afs/cern.ch/user/s/scottw/gangadir/workspace/Local/58/output/stdout

CNTRL D (to quit ganga!)

myscript.sh running...----------------------wn59.itep.ru---------------------- ….. etc.

q (to quit less command)

e.g.HOSTNAME=wn59.itep.ruHOSTNAME=r4b-s16.localHOSTNAME=wn60.itep.ruHOSTNAME=wn072.athena.hellasgrid.gr …. etc.

It seems that you run Ganga for the first time…creates /afs/cern.ch/user/s/scottw/.gangarc

ganga --guito use the gui

Page 15: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

dq2_ls -g user.billscott.testregister011

dq2_ls user.billscott.testregister011

NEED TO REGISTER FILES IN DON-QUIJOTE:

Now register the file(s) in Don Quijote:

nsls -l /castor/cern.ch/user/s/scottw/gjab02/csc.018200.B_Kstar_Kplus_PiMinus_gamma4.evgen._00011.pool.root

List my input file in castor:

pool_insertFileToCatalog rfio:/castor/cern.ch/user/s/scottw/gjab02/ csc.018200.B_Kstar_Kplus_PiMinus_gamma4.evgen._00011.pool.root

Insert the file into the PoolFile Catalog:

dq2_put -p PoolFileCatalog.xml user.billscott.testregister011

List the dataset in Don Quijote:

List the dataset and the files comprising:

(dq2_ls –h for help on dq2_ls)

(dataset name)

(file name)

Page 16: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

j=Job()j.application=AthenaMC()j.application.random_seed='1102362401'j.application.production_name='csc'j.application.process_name='B_Kstar_Kplus_PiMinus_gamma4'j.application.run_number='018200'j.application.firstevent='1'j.application.number_events_job='10'j.application.atlas_release='12.0.6'j.application.transform_archive='AtlasProduction_12_0_6_1_noarch.tar.gz'j.application.mode='simul'j.application.geometryTag = 'ATLAS-CSC-01-02-00' j.application.datasets.indirectory="user.billscott.testregister001"j.application.datasets.inputfile="csc.018200.B_Kstar…gamma4.evgen._00011.pool.root"j.backend=LCG()#j.backend.CE='heplnx206.pp.rl.ac.uk:2119/jobmanager-lcgpbs-atlas'j.backend.CE='ce102.cern.ch:2119/jobmanager-lcglsf-grid_2nh_atlas'j.backend.middleware='EDG'j.outputdata=DQ2OutputDataset()j.application.se_name='RALDISK'j.splitter=AthenaMCSplitterJob()j.splitter.numsubjobs = 2j.splitter.nsubjobs_inputfile = j.splitter.numsubjobs #j.submit()

MY CURRENT GANGA SIMUL/DIGIT JOB:

emacs gsim006.py

Page 17: GETTING STARTED ON THE GRID: W. G. SCOTT (RAL/PPD) RAL PHYSICS MEETING TUES 15 MAY 2007. GENERATED 10K SAMPLES IN EACH CHANNEL ON LXPLUS (IN 2006) SIMULATED/DIGITISDED.

cd /tmpdq2_get -r user.billscott.testregister001 csc.18200.B_Kstar…gamma4.evgen._00011.ntup.root.61

Use Don-Quijote to “Get” the Output file:

WHEN THE JOB HAS RUN:

Remove the job number appended the end of the output filebefore using root:

mv csc.18200.B_Kstar…gamma4.evgen._00011.ntup.root.61 csc.18200.B_Kstar…gamma4.evgen._00011.ntup.rootroot csc.18200.B_Kstar…gamma4.evgen._00011.ntup.root

How do I clean-up/remove all the old jobs/files in Ganga and start fresh??

BIG BROTHER IS WATCHING YOU/ME!!

!less /afs/cern.ch/user/s/scottw/gangadir/workspace/Local/27/0/output/stdout

Use the “less” command to view the Output file:Subjob #Job #

unix command prefixed by shriek !