Chris's mails and MC files

[ytatus94]

• 20170609 /raid05/users/walker/Abe.ark/Abe.01/NUHM2.save/user.walker.TestJob_strong.pool.root

[ytatus94]

• 20170610 /afs/cern.ch/user/w/walker/work/public/Abe/weak/user.walker.TestJob_weak.pool.root or /afs/cern.ch/user/p/pskubic/work/Abe/Abe/weak/user.walker.TestJob_weak.pool.root This is a m1/2 = 350 Weak production file with filter. we think 1000 events.

[ytatus94]

• 20170611 Normally, the job-options (control files) are kept and retrieved from cvmfs to comply with the atlas standard, e.g.:

  data=/cvmfs/atlas.cern.ch/repo/sw/Generators/MC15JobOptions/latest/share/DSID370xxx
  export DATAPATH=${data}:$DATAPATH

  However, we don't have permission to modify these files. So when testing, I make a copy of the MCJobOptions directory and place it in the Abe folder before setting up athena. Furthermore, I keep a local copy of our specific files in my week.600 folder ( .600 is new) and make the changes necessary for these to work from here. This adds an additional point of possible failure, but I find it preferable to trying to change our files under the local copy of the MCJobOptions directory. Later, I will need to test this with our files in the local MCJobOptions directory before we request that they be placed in the standard atlas location.

[ytatus94]

• 20170612
  • Pat /afs/cern.ch/user/p/pskubic/work/Abe/Abe/weak.patest11_600_compressed_success_1000evts/user.pskubic.TestJob_weak.pool.root I have a file (EVNT format) with compressed compressed, and compressed compressed j, events with no filter and m1/2=600. 1000 events, 32MB
  • Pat /raid05/users/walker/Abe.01/test_2.58k_events.weak.600/user.walker.TestJob_weak_600_.pool.root /raid05/users/walker/Abe.01/test_2.58k_events.weak.600/DAOD_TRUTH3.tutorial.pool.root They should have over 5k events with a 2L2 filter. I found the filter eff. was 5.3775% (5915/109995). The filter required at least 2 leptons, pt>2 GeV, eta<2.7, VisPtHadTau>15 GeV

[ytatus94]

• 20170613 /afs/cern.ch/user/w/walker/work/public/Abe.01/test_3.100k_events.filtered.weak.600/user.walker.TestJob_weak_600_.pool.root a filtered run with 100k events

[ytatus94]

• 20170614

  • the 100k events unfiltered run, and generated the truth ntuple for both the filtered and unfiltered. /afs/cern.ch/user/w/walker/work/public/Abe.01/test_3.100k_events.filtered.weak.600 /afs/cern.ch/user/w/walker/work/public/Abe.01/test_4.100k_events.unfiltered.weak.600

  • Okay, they're both running now. I'm not sure how long they will take, but they can be monitored here if you ssh to ouhep05.nhn.ou.edu:

    tail -f  /raid05/users/walker/Abe.01/test_5.100k_events.unfiltered.weak.600

    and

    tail -f  /raid05/users/walker/Abe.01/test_6.100k_events.filtered.weak.600

  • Oops, I left off the nohup.out file name in the previous email (below): More specifically:

    tail -f /raid05/users/walker/Abe.01/test_5.100k_events.unfiltered.weak.600/nohup.out

    and

    tail -f /raid05/users/walker/Abe.01/test_6.100k_events.filtered.weak.600/nohup.out

    see /raid05/users/walker/Abe.01/test_6.100k_events.filtered.weak.600/bkg.sh to see the nohup script. This script is executed from the test...week.600 subdirectory with the following command after modifications to the control files copied from Abe.01/600: nohup ../bkg.sh&

[ytatus94]

• 20170615 /afs/cern.ch/work/w/walker/public/Abe.01/test_5.100k_events.unfiltered.weak.600 /afs/cern.ch/work/w/walker/public/Abe.01/test_6.100k_events.filtered.weak.600 100k events with compressed+compressed and compressed+compressed+j

[ytatus94]

• 20170616 The process we discussed today has completed; 10k events, compressed only with no x1 in the final state, and filtered. This is dubbed test_7. I am now rsync-ing it to my work area underAbe.01. The truth process has been run as well. You can ignore test_8, this was a parallel run with the same parameters and the truth ntuple has not been extracted. Please let me know as soon as possible if you would like me to run a 100k version of this overnight.

[ytatus94]

• 20170619 The job we started has completed. It's located in test_10... of my work area on lxplus. I believe that this has the mass change that we made to w1ss plus compressed only with C1 specified in the final state.

[ytatus94]

• 20170620 Today's test 12 and test 13 have completed and have been rsync't to my work area on lxplus. Test 13 reported a lot of warnings when converting to the truth ntuple. You might want to check the log files to see if everything looks okay.

[ytatus94]

• 20170621 I finally succeeded in running the MC generator using AtlasProduction,19.2.5.21. This is located under test_9... in my work area on lxplus.

The trick was in putting the default card files under a directory called "Cards" and making a few related changes to the control file. See directory "t9" for the initial slate.

I plan to test today with version= 20... as specified in the twiki tutorial. I suspect that this will work with the same changes.

It would help me if Yu-Ting could give test_9... a quick look to make sure it looks okay.

[ytatus94]

• 20170622 Thank you for testing this. Turns out with version 20 my job crashes after MadGraph when initializing Pythia8. I'm still looking in to it.

[ytatus94]

• 20170626
  • Sorry, for some unknown reason I didn't receive your email until now. I'll get started on it right away.
  • Yes, it seems to fail for me too. I suspect that we need a definition for l+ and l-, but I don't really know. I'll try to find why my process fails.
  • Never mind, I found the problem. I had a typo in the control file. It's running now on ouhep05.nhn.ou.edu.
  • Looks like it finished. I'll generate the truth information and resync it to lxplus next. It will be about thirty minutes before I get started on this.
  • Okay, the results are located in my lxplus work area under public/Abe.01/test_15.n2_specific. It took about an hour and 40 minutes to run. Please let me know if the results look okay.

[ytatus94]

• 20170713
  • Thanks, I'll perform the 100k diagnostic with METx10 and METx50, but will also process the 20k with the METx10 multiplier.
  • Thank you, I've just copied the files which I expect contains the latest changes and I'm getting ready to kick off the processes now.
  • I've noticed the code in the Mad...py scripts to multiply the evt_multiplier by 10.0 if doMETFilter is true. However, I've noticed that there is additional code just before this to multiply the evt_multiplier by 100.0 if the 2LFilter is true. Is this what we expect to be doing? i.e. use an evt_multiplier of 1000.0?
  • Okay, I have eight processes running with 100k events each. These are C1C1, N2C1m, N2C1p, and N2N1 for METx10 and METx50. I expect that I should run another eight jobs with 10k events each.
  • Okay, I've kicked off four more processes; METx10 with 20k events.
  • Yes, I noticed the email/JIRA exchange. Good thing we found the problem. Thank you for making the updates. I'll go ahead and download the latest files and have a look.