Project

General

Profile

R009 » History » Version 5

« Previous - Version 5/9 (diff) - Next » - Current version
Kenichi Nakano, 11/19/2018 06:53 PM


R009

Last update on 2018-11-19 by Kenichi.

This production will include the FY2017 dataset and be processed with our local computers.

Some info in this page is not correct for now. Being written...

Information on R009 Production

  • The good runs are listed in "/data2/production/list/R009/R009_fy2017.list".
    • The reasons for good/bad schemas are described in "00readme.txt".
    • The format (columns) of this list file is a bit different from R008 etc.
  • Decoder
    • git hash
  • Data quality check ... "Spill.dataQuality" field
    • Not done yet.
    • The criteria is identical to R007. Details can be found in the Git repository (spillDQRunSchema.py, 606a858a451be9a585ec86ce0434221a812ef7d9).
    • The set of "good" spills is different from R007 because of a requirement of "N of tracks/spill > 0.
    • The "Spill.dataQuality" field was set accordingly on May 17.
    • Analyzer can use the lists of good spills in "/data2/production/list/R008/good_spill_*.list" instead.
  • Chamber intensity ... "QIE.Intensity(_p)" field
  • kTracker
    • Not done yet.
    • The version used is stored in the "kInfo" table.

Details about use of our local computers

  • Computers used
    • e906-gat2, e906-gat4, e906-gat6 (for now?)
  • Decoder parameters
    • Event sampling = 1 (except FPGA2), FPGA2-event sampling = 100
  • Procedure
    • run-local-production.sh
    • check-schema-size.sh
    • summarize-local-production.cc
  • Test #1 - 2018-11-20
    • Runs used ... 299 runs in total
      • Runs 24172-24296 ... first 100 runs of FY2017 dataset
      • Runs 25781-25888 ... middle 100 runs
      • Runs 28612-28740 ... last 99 runs
      • Listed in /data2/production/list/R009/test_local/test_local_gat*.list
    • The decoder exit status of all the runs is OK (i.e. has reached the end of the program).
    • Statistics
      • The averaged process time = 6149 sec/run ... The process time per input Coda size (sec/MB) is shown in the attachment "h1_t2si.png". The three peaks correspond to the three run groups (or servers).
      • Input Coda size = 3143 MB/run on average
      • Output ROOT size = 637 MB/run on average
      • Output / input size = 0.20 on average ... The distribution is shown in the attachment "h1_so2si.png". The three peaks correspond to the three run groups. The outliers might be problematic.

Reference

  • docdb 4159: Estimate of output size etc.
  • docdb 4242: FPGA2-event sampling