CMSSW Mixing Module - Known issues

The modul loaded by the --pileup AVE_35_BX_25ns cmsDriver.py option:

process.load('SimGeneral.MixingModule.mix_POISSON_average_cfi')

http://cmslxr.fnal.gov/source/SimGeneral/MixingModule/python/mix_POISSON_average_cfi.py

Extra commands in the generated cfg file (these are implicit in the flat configuration):

process.mix.bunchspace = cms.int32(25)
process.mix.minBunch = cms.int32(-12)
process.mix.maxBunch = cms.int32(3)

The modul loaded by the --pileup Flat_20_50 cmsDriver.py option:

process.load('SimGeneral.MixingModule.mix_Flat_20_50_cfi')

http://cmslxr.fnal.gov/source/SimGeneral/MixingModule/python/mix_Flat_20_50_cfi.py

The differecences between the two mixing modules (extracted using the python interactive mode) (diff -ub poisson flat):

>         OOT_type = cms.untracked.string('Poisson'),                        # property of process.mix.input

>         manage_OOT = cms.untracked.bool(True),                             # property of process.mix.input

<             averageNumber = cms.double(35.0)                               # property of process.mix.input.nbPileupEvents

---

>             probFunctionVariable = cms.vint32(0, 1, 2, 3, 4,               # property of process.mix.input.nbPileupEvents

>                 5, 6, 7, 8, 9, 

>                 10, 11, 12, 13, 14, 

>                 15, 16, 17, 18, 19, 

>                 20, 21, 22, 23, 24, 

>                 25, 26, 27, 28, 29, 

>                 30, 31, 32, 33, 34, 

>                 35, 36, 37, 38, 39, 

>                 40, 41, 42, 43, 44, 

>                 45, 46, 47, 48, 49, 

>                 50),

>             probValue = cms.vdouble(0, 0, 0, 0, 0,                         # property of process.mix.input.nbPileupEvents

>                 0, 0, 0, 0, 0, 

>                 0, 0, 0, 0, 0, 

>                 0, 0, 0, 0, 0, 

>                 0.032258065, 0.032258065, 0.032258065, 0.032258065, 0.032258065, 

>                 0.032258065, 0.032258065, 0.032258065, 0.032258065, 0.032258065, 

>                 0.032258065, 0.032258065, 0.032258065, 0.032258065, 0.032258065, 

>                 0.032258065, 0.032258065, 0.032258065, 0.032258065, 0.032258065, 

>                 0.032258065, 0.032258065, 0.032258065, 0.032258065, 0.032258065, 

>                 0.032258065, 0.032258065, 0.032258065, 0.032258065, 0.032258065, 

>                 0.032258065)

<         type = cms.string('poisson')                                       # property of process.mix.input

---

>         type = cms.string('probFunction')                                  # property of process.mix.input

process.mix.input.OOT_type = cms.untracked.string('Poisson')
process.mix.input.manage_OOT = cms.untracked.bool(True)

Adding this lines to a cmsDriver configuration file still seems to produce errors.

Known error messages:

Message:

OscarMTMasterThread::stopThread() has not been called to stop Geant4 and join the master thread
terminate called without an active exception

This is an useless error message. It is known to occur when the probability function used to generate the pileup distribution is not properly normalized. A guess for what it might mean: The mixingmodule starts children processes to generate a given number of one-pileup events at the same time for having them mixed later. Whenever it requires no more pileups mixed, it stops waiting for its children processes and it might cause some kind of error.

Message:

----- Begin Fatal Exception 17-Feb-2017 11:15:31 CET-----------------------
An exception of category 'FatalRootError' occurred while
   [0] Processing run: 1 lumi: 1 event: 1
   [1] Running path 'HLTriggerFirstPath'
   [2] Calling event method for module HLTGetRaw/'hltGetRaw'
   [3] Calling produce method for unscheduled module RawDataCollectorByLabel/'rawDataCollector'
   [4] Calling produce method for unscheduled module SiStripDigiToRawModule/'SiStripDigiToRaw'
   [5] Calling produce method for unscheduled module MixingModule/'mix'
   Additional Info:
      [a] Fatal Root Error: @SUB=TH1F::ComputeIntegral
Integral = zero

----- End Fatal Exception -------------------------------------------------

This error seems to come from the pileup mixing, possibly from a bad normalization: http://cmslxr.fnal.gov/source/Mixing/Base/src/PileUp.cc

-- AdamHunyadi - 2017-02-21