Tags:
, view all tags
-- BrendanKiburg - 15 Jan 2008

The following are some of the tasks that should be in place before starting the mini production pass in the next week or so

Update 20 Feb

  1. Processed a few hundred runs. Ran into problems with the mta analysis seg faulting on some fraction of the files
  2. Investigated seg faults. Problem had to do with multiple methods trying to release memory
  3. Restarted Last night. More files are processed, but seg faults remain
  4. In general, you can find the files at /usr/projects/bkiburg/run10/pass3_prod/merged_ana.root or merged_mta.root for investigation for run10, and /usr/projects/bkiburg/run11/pass3_prod/ for run11

Update 13 Feb

  1. DONE Seg. Faults were generated by MIDAS buffer overruns. MIDAS buffer have been increased and recompiled in
  2. DONE EH Pixel Selection definition were refined -
  3. DONEMusc_lo has been added to an additional set of banks, available in mu, allowing an out of time musc_lo to veto an otherwise good muon
  4. HourglassNeed to add a flag to the tree when the a musc-hit would be disallowed due to the the musclo veto
  5. After this mini-pass, we will include some of Steven's code (Peter has worked on this already)
  6. Plot updates
    1. How much does musc-lo vetoing suppress the data rate : (~8%). Red curve is with musclo, Black is traditional one. musc_pugate.png
    1. How does EH anode requirement affect run->run conclusions?
      Fixed definition wrt 0 bin - added GoodMuon /BestEntrance requirement (in red)
      contEH.png
    1. Reran on run10 and run11 to show run11 has more anodes, on average. Results not ready yet runcomp.png

New Issues To Address

  1. People should sign off on Neutron/WFD/Fadc analysis
  2. Insert a check into the analysis that the run number in the odb matches the run number in the file name
  3. DONEUpdate earlier /src/shared modules to take advantage of run9 run10 flexibility in timing offsets [This was a large step]
    1. DONEUpdate the Runinfo/Run number variable earlier in the analysis (mana.c) so that this is available at mehod initialization
    2. Several parameters have been modified to have sub branches in the odb, based on run number
    3. A new master.odb file should be used with all analysis (CVS is up-to-date)
  4. DONEModify run list for run10 to focus on runs with higher quality FADC data (avoid 385xx and use 423xx instead) [Note: some files won't have full FADC updates, but the original list has a good selection from throughout the run]
  5. DONERemove unnecessary diagnostic output from the main
  6. Peter would like to incorporate some of Steven's MucapDevel code into the production. We need to see how many differences there are in his most up-to-date version and how this may conflict with our more recent code developments
  7. HourglassUpdate the mta analysis to provide more useful output
    1. EH Anodes at end of track (total vs. continuous) EH_eot.png
    2. Lifetime vs. pixels tdist.png
    3. Entrance counter plots
  8. DONE Fixed some memory leaks in the code
  9. Recent investigation in the mta analysis has produced occasional seg-faults. The frequency and or causes need to be diagnosed before producing lots of trees that cannot be processed in the subsequent pass

Analysis information

  1. DONECVS should be kept up to date so developers can properly integrate their code
  2. DONEFADC analysis integrated into the production code
  3. DONENeutron analysis integrated into the production code
  4. DONEWFD analysis integrated into the production code
  5. Stop Discussion about what data, if any, should be skimmed from these files :next pass
  6. My changes with respect to : entrance counters, MuStop defintion
Organizational information
  1. CVS should be kept up to date so developers can properly integrate their code
  2. Hourglass Project space needs to be allocated from NCSA : in progress
  3. DONE ODB files and Parameters file need to be used separately for run 10, run 11 files
  4. DONE A run list needs to be selected (neutron, muminus, muplus, run10, run11)
    1. Proposed run list for run 10 files minipass_run10_files.txt (10 neutron, 100 mu plus, 500 mu minus)
    2. Proposed run list for run 11 files minipass_run11_files.txt ( 39 neutron, 200 argon, 100 muplus, 200 mu minus)
  5. DONEThe files need to be transferred to scratch or project space prior to launching the jobs. Run 10 files took about 26 hours for ~700 files.

Proposed plan

  1. "Complete" the steps in "Analysis Information," and run over a few files
  2. Individual users check that relevant histograms/data appear in the root/tree files and match the output from their code
  3. When those are signed off on, we can run over several hundred files with 'mu'
  4. Those can be processed with mta
  5. Results can be summed and examined

Full file listings

Topic attachments
I Attachment History Action Size Date Who Comment
PNGpng EH_eot.png r1 manage 8.9 K 2008-02-06 - 06:35 BrendanKiburg EH Anodes at end of track
Texttxt argon.txt r1 manage 3.3 K 2008-01-18 - 19:41 BrendanKiburg Argon_run11.txt
PNGpng contEH.png r1 manage 8.8 K 2008-02-13 - 06:55 BrendanKiburg cont EH pixels with good muon def
Texttxt golden_minus_run10.txt r1 manage 33.9 K 2008-01-18 - 19:44 BrendanKiburg Golden_muminus_run10.txt
Texttxt golden_plus_run10.txt r1 manage 8.3 K 2008-01-18 - 19:44 BrendanKiburg Golden_muplus_run10.txt
Texttxt minipass_run10_files.txt r1 manage 3.6 K 2008-01-18 - 20:03 BrendanKiburg Mini Pass Run 10 files
Texttxt minipass_run11_files.txt r1 manage 3.0 K 2008-01-18 - 20:17 BrendanKiburg Mini Pass Run 11 files
PNGpng musc_pugate.png r1 manage 9.2 K 2008-02-13 - 06:33 BrendanKiburg musc_pugate_def
Texttxt neutron_run10.txt r1 manage 0.1 K 2008-01-18 - 19:41 BrendanKiburg Neutron_run10.txt
Texttxt neutron_run11.txt r1 manage 0.2 K 2008-01-18 - 19:42 BrendanKiburg Neutron_run11.txt
PNGpng runcomp.png r1 manage 9.0 K 2008-02-13 - 15:40 BrendanKiburg EH run comparison
PNGpng tdist.png r1 manage 37.1 K 2008-02-06 - 06:36 BrendanKiburg Time Dist vs. Anode info
Edit | Attach | Watch | Print version | History: r14 < r13 < r12 < r11 < r10 | Backlinks | Raw View | Raw edit | More topic actions...
Topic revision: r13 - 2008-02-20 - BrendanKiburg
 
  • Edit
  • Attach
This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback