Difference: MiniProd2Run10 (1 vs. 14)

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. Seg. Faults were generated by MIDAS buffer overruns. MIDAS buffer have been increased and recompiled in
2. EH Pixel Selection definition were refined -
3. Musc_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. Need 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.
7. 1. How does EH anode requirement affect run->run conclusions?
      Fixed definition wrt 0 bin - added GoodMuon /BestEntrance requirement (in red)
      
8. 1. Reran on run10 and run11 to show run11 has more anodes, on average. Results not ready yet

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. Update earlier /src/shared modules to take advantage of run9 run10 flexibility in timing offsets [This was a large step]
   1. Update 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. Modify 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. Remove 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. Update the mta analysis to provide more useful output
   1. EH Anodes at end of track (total vs. continuous)
      
   2. Lifetime vs. pixels
      
   3. Entrance counter plots
8. 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. CVS should be kept up to date so developers can properly integrate their code
2. FADC analysis integrated into the production code
3. Neutron analysis integrated into the production code
4. WFD analysis integrated into the production code
5. Discussion about what data, if any, should be skimmed from these files :next pass
6. My changes with respect to : entrance counters, MuStop defintion

1. CVS should be kept up to date so developers can properly integrate their code
2. Project space needs to be allocated from NCSA : in progress
3. ODB files and Parameters file need to be used separately for run 10, run 11 files
4. 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. The 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

• argon.txt: Argon_run11.txt

• neutron_run10.txt: Neutron_run10.txt

• neutron_run11.txt: Neutron_run11.txt

• golden_plus_run10.txt: Golden_muplus_run10.txt

• golden_minus_run10.txt: Golden_muminus_run10.txt

1. 1. How does EH anode requirement affect run->run conclusions?
      Fixed definition wrt 0 bin - added GoodMuon/BestEntrance requirement (in red)
      

1. 1. EH Anodes at end of track (total vs. continuous)
      
   2. Lifetime vs. pixels
      

1. Update the Runinfo/Run number variable earlier in the analysis (mana.c) so that this is available at mehod initialization

1. Update earlier /src/shared modules to take advantage of run9 run10 flexibility in timing offsets

1. Remove unnecessary diagnostic output from the main analysis
2. Update the mta analysis to provide more useful output

1. The files need to be transferred to scratch or project space prior to launching the jobs.
   

1. Update earlier /src/shared modules to take advantage of run9 run10 flexibility in timing offsets
2. Modify run list for run10 to focus on runs with higher quality FADC data (avoid 385xx and use 423xx instead)
3. Remove unnecessary diagnostic output from the main analysis

1. ODB files and Parameters file need to be used separately for run 10, run 11 files

1. Neutron analysis integrated into the production code

1. WFD analysis integrated into the production code

1. WFD analysis integrated into the production code

1. CVS should be kept up to date so developers can properly integrate their code
2. FADC analysis integrated into the production code

1. Discussion about what data, if any, should be skimmed from these files next pass

1. Project space needs to be allocated from NCSA : in progress

1. A run list needs to be selected (neutron, muminus, muplus, run10, run11)

1. Discussion about what data, if any, should be skimmed from these files [This only needs to be done once]

1. Project space needs to be allocated from NCSA