Tighten Ds*, K*Gamma and PhiGamma and add new line Eta2MuMuGamma

I tried the following modifications to the CaloPID lines:

• include the missing stripping cuts on the Ds, for that I added a new Filter that applies to the Hlt2Shared_Ds;

• use TightKstar2KPi and UnbiasedPhi2KK in place of Kstar2KPi and Phi2KK for the KGamma and PhiGamma lines. In that way, I only modified Inputs.py and not the Combiners in Stages.py. The following additional cuts are now applied: PIDK(K)>0, PT(K+,pi-)>500 MeV and mass window from 300 to 100 MeV for the K and from 50 to 30 for the Phi. Looking at the offline cuts used for the Run1 PID samples, these are all used, so I think it is OK.

The Moore rate test results using v26r5p1 are:

• from 30 to 10 Hz for the Ds* line;

• from 10 to <10 Hz for the K*Gamma;

• from <10 to <10 Hz for the PhiGamma;

So we saved 20-30 Hz which can be used to either lower the PT cut of the Ds* line or include the Eta2MuMuGamma line. (RM: the latter was done)

Log files with old and new settings are attached.

moore_oldcuts.log

moore_newcuts.log

I tried the following modifications to the CaloPID lines:

• include the missing stripping cuts on the Ds, for that I added a new Filter that applies to the Hlt2Shared_Ds;

• use TightKstar2KPi and UnbiasedPhi2KK in place of Kstar2KPi and Phi2KK for the KGamma and PhiGamma lines. In that way, I only modified Inputs.py and not the Combiners in Stages.py. The following additional cuts are now applied: PIDK(K)>0, PT(K+,pi-)>500 MeV and mass window from 300 to 100 MeV for the K and from 50 to 30 for the Phi. Looking at the offline cuts used for the Run1 PID samples, these are all used, so I think it is OK.

For some reasons, I need to include the Exotica module in Physics_pp_2017 for Moore to run. Running with CaloPID only does not work.

I was thinking about it, but these are the cuts of the D2hhhFTCalib_KKPLine used in run2 strippings which is a standard object. Are we really ever want to change them? Or if it is more a matter of writing rules, I'll stick to the rules.

Then we might want to move it to Hlt2SharedParticles but I leave that question to the Hlt team :) The rest LGTM.

About the issue with the rate test, I did some checks and my conclusion is that the problem arises when the test is run with NO PersistReco line. As soon as you add one (Exotica is an small module with at leat one PersistReco line) it runs fine. This happens also with Moore v25r4 but did not happen with Moore v25r1. For completeness, the error one gets if running only the CaloPID module is:

HltPersistRecoClonerSequence               INFO Member list: GaudiSequencer/HltFullPersistRecoSequence, Gaudi
Sequencer/HltCopyTurboOutputsSequence
JobOptionsSvc                             ERROR Unable to set the property 'Members' of 'HltFullPersistRecoSequence'. Check option and algorithm names, type and bounds.
HltPersistRecoClonerSequence              ERROR GaudiSequencer:: Can not initialize HltFullPersistRecoSequence StatusCode=FAILURE
HltPersistRecoSequence                    ERROR GaudiSequencer:: Can not initialize HltPersistRecoClonerSequence StatusCode=FAILURE
HltAfterburnerSequence                    ERROR GaudiSequencer:: Can not initialize HltPersistRecoSequence StatusCode=FAILURE
HltAfterburnerFilterSequence              ERROR GaudiSequencer:: Can not initialize HltAfterburnerSequence StatusCode=FAILURE
HltAfterburner                            ERROR GaudiSequencer:: Can not initialize HltAfterburnerFilterSequence StatusCode=FAILURE
HltDecisionSequence                       ERROR GaudiSequencer:: Can not initialize HltAfterburner StatusCode=FAILURE
Hlt                                       ERROR GaudiSequencer:: Can not initialize HltDecisionSequence StatusCode=FAILURE
EventLoopMgr                              ERROR Unable to initialize Algorithm: Hlt
ServiceManager                            ERROR Unable to initialize Service: EventLoopMgr
ServiceManager                            ERROR Service EventLoopMgr not in the correct state to be initialized (CONFIGURED)
ServiceManager                            ERROR Service EventLoopMgr not in the correct state to be initialized (CONFIGURED)
ApplicationMgr                            FATAL stop: Invalid state "CONFIGURED"

• [2017-08-26 00:07] Validation started with lhcb-2017-patches#121
• [2017-08-27 00:06] Validation started with lhcb-2017-patches#122
• [2017-08-28 00:06] Validation started with lhcb-2017-patches#123
• [2017-08-29 00:05] Validation started with lhcb-2017-patches#124
• [2017-08-30 00:06] Validation started with lhcb-2017-patches#125
• [2017-08-31 00:05] Validation started with lhcb-2017-patches#126
• [2017-09-01 00:05] Validation started with lhcb-2017-patches#127
• [2017-09-02 00:06] Validation started with lhcb-2017-patches#128
• [2017-09-03 00:07] Validation started with lhcb-2017-patches#129
• [2017-09-04 00:07] Validation started with lhcb-2017-patches#130

added 1 commit

• 1dc828e6 - Add new neutral line Eta2MuMuGamma

Compare with previous version

New TURCAL line using DiMuon trigger. Largely based on Hlt2ExoticaPrmptDiMuonTurbo:

• DiMuonNoIPCombiner;
• PrmptDiMuonFilter;

then add gamma with PT>500 MeV:

• Eta2MuMuGammaCombiner.

Rate tests on 100k events:

• 30 Hz with PT>200 MeV (test_eta2mumugamma_100k_200mev.log);
• 20 Hz with PT>500 MeV (test_eta2mumugamma_100k_500mev.log).

changed title from {-Add missing cuts to the Ds used for the Ds - tighten K*gamma and Phigamma cuts-}* to {+Tighten Ds, K*Gamma and PhiGamma and add new line Eta2MuMuGamma+}*

changed the description

added ~1529 line ~1530 labels

changed milestone to %v26r6

assigned to @rmatev

resolved all discussions

approved this merge request

merged

mentioned in commit e16771f7

@rmatev HltSettings needs to be updated to include the new Eta2MuMuGamma line and cuts. @chefdevi was working on it but I'm afraid he couldn't finish it before going on holidays. When do you want to deploy the new version? I could update the settings if he is not back by then.

Oops, I didn't see that. He must have done it for his tests. Any chance you can do it this afternoon?

yes, he had problems committing it to git. I'm on it!