HLT1 Lumi Line

First version of Lumi Line

Resolve the Allen algorithms part of the open issue #220

Extend the hlt1_pp_default sequence with the following algorithms:

• CalcLumiSumSize

  • Read the decision of HltODINLumiLine, if the event fired the trigger, set the corresponding lumi_summary_size
    • The size is defined in device/lumi/include/LumiCounterOffsets.h

• MakeLumiSummary

  • Write the Lumi counters by a specific order into the HLTLumiSummary rawbank, with data type unsigned, and assign it to HostBuffer
    • Store multiple counters in one unsigned int
    • The order, bit offsets, and sizes are defined in the header file device/lumi/include/LumiCounterOffsets.h
  • Also with necessary ODIN information for lumi analysis
    • see discussion in twiki page DiscussionOnLumiEvents
    • event time defined as: t0 = gps - (orbit*3564 + bcid) *1000 / 40078
      • The event time is not rounded for simplicity
    • re-mapped bunch id defined as: bcid = orbit*3564 + bcid
    • bunch crossing type
  • Current counters:
    • passed GEC or not
    • the number of Velo tracks
    • the number of Velo Vertices
    • the number of SciFi clusters
    • Muon hits of region
      • M2R2
      • M2R3
      • M3R2
      • M3R3

Lumi Line uses PVs as lumi counters and requires running the Velo without GEC. This introduces a ~5% slowdown.

Update the reference files since the ordering that the MES (Multi Event Scheduler) gives is changed.

Add Rec/Allen/src/GaudiAllenLumiSummaryToRawEvent to support Moore-based Gaudi Allen configuration

• copied from Rec/Allen/src/GaudiAllenReportsToRawEvent

Merged from !765 (closed), which add lumi_summaries and lumi_summary_offsets to HostBuffer

• Also added lumi summaries to OutputHandler.
• The size of lumi summaries is expected to be much smaller than sel reports.

added Luminosity label

added RTA label

@dcraik Hi, Dan. I made a to-do list. Since I'm totally new to HLT1, could you please have a check? If the things listed here are already done or anything is not right, please tell me. Thanks in advance.

Lumi counters to be reconstructed and write to HLTLumiSummary:

See https://codimd.web.cern.ch/SMAYY_DRSDOsDGLKhzPf7w

Velo discussion

SciFi discussion

Calo discussion

Muon discussion

added 1 commit

• d00c3f15 - create HLTLumiSummary.cuh

Compare with previous version

changed the description

changed the description

changed the description

added 1 commit

• a64998c8 - add device algorithm `make_lumi_summary`

Compare with previous version

mentioned in merge request !765 (closed)

I've created a branch with the writing apparatus here: !765 (closed)

Your algorithm should write the LumiSummaries and their offsets to the containers in the HostBuffers here: https://gitlab.cern.ch/lhcb/Allen/-/blob/1e821f537d10151de73602fe07ee53dd7d8883ab/stream/sequence/include/HostBuffers.cuh#L134 For events without a LumiSummary, the offset should be the offset of the next object (or else the end marker) such that offset[x+1] - offset[x] = 0 (which will tell the code to skip the LumiSummary for that event).

added 34 commits

• a64998c8...a090c07b - 31 commits from branch master
• 1e821f53 - First version of writing LumiSummary objects to output (if they exist)
• 9936bc60 - Merge remote-tracking branch 'origin/dcraik_lumi_summaries' into shuxian_LumiLine
• c5adce8b - merge codes from Dan on the writing side, fix device algorithm MakeLumiSummary

Compare with previous version

added 1 commit

• 3a7fe78a - add device algorithm CalcLumiSumSize

Compare with previous version

changed the description

changed the description

changed the description

changed the description

added 1 commit

• 8d7f08eb - add device algorithm MakeLumiTrackCounters, and gather all lumi related device...

Compare with previous version

added 1 commit

• 27da0782 - add host algorithm HostLumiEventFilter

Compare with previous version