Conversation
|
Hallo Raphael, currently CMSSW 8 is not covered by the automated travis tests (@greyxray can give more details), but maybe still by Fabio's tests. Only CMSSW 7 is covered by travis. So one might have to perform a manual test here... I do not consider dropping the refitting as a good option since this would mean that Claudia would not be able to share Artus inputs with us but would need a complete own skim. We are more limited by storage space then by computing time. Regards, |
|
Hi Raphael, I would like to keep the RefitVertex code, since it's crucial for my analysis. Thanks. Best, |
|
|
||
| # setup BadPFMuonFilter and BadChargedCandidateFilter | ||
| if tools.is_above_cmssw_version([8]): | ||
| if tools.is_above_cmssw_version([8]) and not tools.is_above_cmssw_version([8]): |
There was a problem hiding this comment.
Could it be that you forgot to change the latter version requirement to 9?
|
|
||
| ## if needed adapt output filename | ||
| process.p *= process.kappaOut | ||
| process.ep *= process.kappaOut |
There was a problem hiding this comment.
Just for my curiosity: what is the difference between process.p and process.ep?
There was a problem hiding this comment.
"p" is the path, "ep" is the endpath. Running Kappa in the path creates a unsolveable ordering problem of the unscheduled mode. By putting it to the endpath, everything is fine.
|
Just FYI: Olena is on holidays for at least a week. So you might have to wait some time for answers regarding the tests. |
This reverts commit 80f6f49.
| process.kappaTuple.TriggerObjectStandalone.metfilterbitslist = cms.vstring("BadChargedCandidateFilter","BadPFMuonFilter") | ||
| # process.kappaTuple.TriggerObjectStandalone.metfilterbitslist = cms.vstring("BadChargedCandidateFilter", "BadPFMuonFilter", "badGlobalMuonTaggerMAOD", "cloneGlobalMuonTaggerMAOD") | ||
| # else: | ||
| # process.kappaTuple.TriggerObjectStandalone.metfilterbitslist = cms.vstring("BadChargedCandidateFilter","BadPFMuonFilter") |
There was a problem hiding this comment.
Sorry for perhaps another stupid question, but do these lines need to be commented?
5 participants
Dear all,
after all a few changes had to be made to get Kappa Running on CMSSW. This is especially due to the TriggerObjectStandalone Collection that had some changes. They should be backward-compatible, but I would leave that to the automated tests complaining.
MVAMET is deactivated since there's no CMSSW9 training yet.
Also, the re-correction of jets is deactivated. As soon as there are relevant new JECs we should change that again. The reason is just that we used a module from the MVAMET implementation to re-correct the jets.
The refitted vertex does run. However if not totally necessary we should think about dropping it: Runtime-Wise it is the dominant module, by far: out of 0.166 s/event 0.15 s/event are due to the AdvancedRefitVertexBSProducer and AdvancedRefitVertexNoBSProducer producers on a 2017B SingleElectron Data file.
@greyxray is it a lot of work to extend the automated tests to the 92X release and run the old ones on this branch? I think nothing should change in the old releases, but this has to be proven.