L1Calo Offline Software Status

Transcript

1. L1Calo  Offline  and   Monitoring  Software Alexander  Mazurov (on  behalf  of  many  people) L1Calo  Joint  Meeting,  22  June  2015
2. Content • Athena  releases • Conditions  database • ByteStream  decoders • Calibration • Simulation • Monitoring 22  June  2015 L1Calo  Joint  Meeting 2
3. Athena  Release:  Builds • We  contribute  to  the  following  Athena  builds: • Tier0: AtlasProduction,  20.1.X.Y-­‐VAL • P1: AtlasCAFHLT,  20.2.X.Y.Z-­‐VAL • Edge: AtlasOffline,  20.X.Y-­‐VAL • P1MON is  planned  to  be  build  soon  (end  of  this  week?)  for  shorter   release  cycles  of  monitoring  packages. 22  June  2015 L1Calo  Joint  Meeting 3
4. Athena  Releases:  Tests Don’t  break  releases.  Test,  test,  test!   • Tier0.  Has  frozen  policy:  not  permitted  when  modifications  change  existing   EDM/xAOD  components.  Most  important  when  we  change  event  objects  or   bytestream  decoders.   • https://twiki.cern.ch/twiki/bin/viewauth/Atlas/TriggerAtTier0#Specific_tests • P1  https://twiki.cern.ch/twiki/bin/view/Atlas/HowToSubmitATag#HLT_online_AtlasCAFHLT_cache • These  tests  take  time  to  run,  but  we  could  not  request  our  tags  to  be  collected   without  specifying  what  tests  are  passed. • We  need  to  have  our  L1Calo  regression  tests.  For  example: • Test  that  content  of  L1Calo  xAOD  root  files  does  not  change  after  bytestream  decoders  update • Check  that  number  and  content  of  monitoring  histograms  don’t  changes  in  compression  with   some  reference  histograms 22  June  2015 L1Calo  Joint  Meeting 4
5. Conditions  Database:  Changes • During  LS1  L1Calo  was  using  a  private  SQLite  file  as  COOL  database,   with  various  schema  changes. • All  the  new  folders  have  now  been  created  in  the  CONDBR2  COOL   database  and  the  contents  of  LS1  SQLite  file  have  been  uploaded  to   the  new  folders. • Recent  software  updates  enable  the  access  to  the  latest  Run-­‐2   conditions  database: • Makes  the  monitoring  histograms  that  compares  data  and  simulation  of  the   nMCMs  usable • Fix  calibration  conditions  access 22  June  2015 L1Calo  Joint  Meeting 5
6. Conditions  Database:  Paths • Run-­‐1: • /TRIGGER/L1Calo/V1/Calibration/Calib1/PprChanCalib • /TRIGGER/L1Calo/V1/Calibration/Calib2/PprChanCalib • /TRIGGER/L1Calo/V1/Calibration/Cosmics/PprChanCalib • /TRIGGER/L1Calo/V1/Configuration/ReadoutConfig • /TRIGGER/Receivers/Conditions/VgaDac • /TRIGGER/Receivers/Conditions/Strategy • /TRIGGER/L1Calo/V1/Conditions/RunParameters • /TRIGGER/L1Calo/V1/Conditions/DerivedRunPars • /TRIGGER/L1Calo/V1/Configuration/PprChanDefaults • /TRIGGER/L1Calo/V1/Calibration/Physics/PprChanCalib • /TRIGGER/L1Calo/V1/Calibration/PpmDeadChannels • /TRIGGER/L1Calo/V1/Conditions/DisabledTowers 22  June  2015 L1Calo  Joint  Meeting 6 • Run-­‐2: • /TRIGGER/L1Calo/V2/Calibration/Physics/PprChanCalib • /TRIGGER/L1Calo/V1/Conditions/RunParameters • /TRIGGER/L1Calo/V1/Conditions/DerivedRunPars • /TRIGGER/Receivers/Conditions/VgaDac • /TRIGGER/Receivers/Conditions/Strategy • /TRIGGER/L1Calo/V1/Calibration/PpmDeadChannels • /TRIGGER/L1Calo/V2/Configuration/PprChanDefaults Several  paths  were  updated
7. ByteStream  Decoders • PPM • Recent  change:  set  pedestal  correction  enabling  bit  by  default  – tag  was  recently  accepted  to   tier0  Athena  build • Issue  with  SubBlock  error  handling(ATLLONECAL-­‐14) • Most  of  the  decoders  were  rewritten  to  fill  xAOD  objects.  Some  are  released,  but   some  are  waiting  for  monitoring  and  simulation  to  be  updated: 22  June  2015 L1Calo  Joint  Meeting 7 Released: • TriggerTower • CPMTriggerTower • CMXCPTob • CMXCPHit Ready  but  not  released: • CMXJetHits • CMXJetTob • CMXEtSums • JEMEtSums • CPMTobRoI • JEMTobRoI In  progress: • JetElement • CMXRoi • RODHeader
8. Simulation 22  June  2015 L1Calo  Joint  Meeting 8 Sequence  of  algorithms  and  data  transfers  more  closely  follows  the  hardware  than  in  Run  1
9. Simulation:  Updates • Separate  thresholds  in  CP  and  JEM  LUTs,  adjustments  to  noise  cuts,  BCID  settings &   baseline  subtraction  in  PreProcessorsimulation. • Use  xAOD::CPMTower  as  its  input  to  CPM  simulation. • CPM  LUT  simulation  switched  to  integer  arithmetic  (following  hardware  configuration). • Additions  for  handling  of  restricted  eta  MET/TE  triggers. • Switch  MET,  TE,  XS  triggers  to  emulate  firmware  logic  (arithmetic  based)  rather  than  Run-­‐ 1  (LUT  based).  May  need  further  validation  against  data. • Adjustments  to  TOB  objects:  changes  to  what  data  the  CPTopoTOB  and  JetTopoTOB   objects  can  output. Changes  to  integer  coordinate  convention  of  JetTopoTOB. • Bug  fixes,  coverity  fixes,  updates  to  current  Athena  framework,  etc. 22  June  2015 L1Calo  Joint  Meeting 9
10. Simulation:  Tasks • Some  update  of  Jet  eta  bins  for  menu  needs  implementing  (though   offline  had  fewer  problems  than  original  online).  Need  to  re-­‐confirm   that  objects  used  in  steering  are  OK  with  the  new  scheme. • Switch  simulation  to  use  xAOD::JetElement  instead  of   LVL1::JetElement,  to  be  coordinated  with  update  to  BS  decoders  and   monitoring.   • Still  some  aspect  of  the  restricted  eta  ET triggers  that  need  updating.   • Migration  of  some  more  outputs  to  xAOD. 22  June  2015 L1Calo  Joint  Meeting 10
11. Calibration • RampMaker was  updated  to  use  new  xAOD  trigger  towers  objects. • No  evident  problems  with  Tile calibration  tools • LAr calibration:   • Fixed  a  problem  with  missing  timestamps  used  for  database  access. • This  is  a  dirty  workaround  and  the  underlying   problem  should   be  cured. • Issue  with  LAr  conditions  reading  for  HEC  and FCAL calibration.  It  could  be  the  same   timestamp  problem. • Was  updated  an  algorithm  that  creates  a  SQLite  database  of  current   HV  corrections. • Calibration  results  at  Juraj’s talk  tomorrow 22  June  2015 L1Calo  Joint  Meeting 11
12. Monitoring:  PPM PPM  monitoring  was  already  in  good  shapes  during    the  last  L1Calo  Joint   meeting.   Updates: • 2D-­‐PedestalCorrectionVsBCIDVsLumiBlock  summary  histograms • Adjusted  the  PPM  monitoring  for  the  new  database Tasks: • Monitoring  of  80MHz  saturated  BCID  for  25ns  operation  mode • Optionally   improve  2D-­‐PedestalCorrectionVsBCIDVsLumiBlock:   automatically  update  the  lumi  block  scale. 22  June  2015 L1Calo  Joint  Meeting 12
13. Monitoring:  CP,  JEP  and  CMX Released: • CP  and    JEP  monitoring • CMX  CP    and  CMX  JEP  monitoring. • Replaced  all  the  CMM  old  histograms  with  the  new  CMX  histograms • Tool  for  comparison  with  simulation  for  the  CP  data Ready,  but  not  released  yet: • Tool  for  comparison  with  simulation  for  the  JEM  data • Simulation  packages  need  to  be  updated  for  the  new  xAOD  objects Tasks: • Migrate  all  monitoring   tools  to  xAOD  objects • Check  and  update  problematic  histograms 22  June  2015 L1Calo  Joint  Meeting 13 A  lot  of  updates!
14. Monitoring:  Weekly  Meeting • We  have  a  weekly  on  Thursday  on  monitoring  tasks.  Currently  it  starts   after  L1Calo  Chat  Meeting  (around  10:30  or  later)  and  It’s  a  good   place  to  discuss  issues  from  the  other  areas  of  L1Calo  Offline   software.  At  these  meeting  we  discuss  our  JIRA  tickets  -­‐ https://its.cern.ch/jira/secure/RapidBoard.jspa?rapidView=3794 • E-­‐group:  atlas-­‐lvl1-­‐calo-­‐monitoring   22  June  2015 L1Calo  Joint  Meeting 14
15. Monitoring:  Summary • A  lot  of  updates  in  CP,  JEP,  CMX  and  ROD  monitoring • New  L1Topo  histograms • Updated  DQ  web  display  configuration • Observations  from  the  monitoring  histograms  at  Kate’s  talk  later   today   22  June  2015 L1Calo  Joint  Meeting 15
16. Summary • A  lot  of  packages  updated  for  simulation,  monitoring  and  calibration  tasks. • Main  changes  and  issues  were  due  to  the  new  object  model  for  trigger   towers  (separate  hadronic  and  em  towers)  and  migration  to  xAOD  objects • Migrated  to  RUN2  COOL  Oracle  database. • Calibration  jobs  work,  but  the  issues  with  LAr  condition  database  access   need  to  be  resolved.   • We  need  to  check  all  monitoring  histograms  and  fix  the  raised  issues. • Updated  bytestream  decoders  for  the  new  xAOD  model.  Many  tools  were   migrated,  bit  we  need  to  update  the  rest  simulation  and  monitoring  tools   (urgent!). 22  June  2015 L1Calo  Joint  Meeting 16