athena merge requestshttps://gitlab.cern.ch/atlas/athena/-/merge_requests2019-05-08T03:02:59+02:00https://gitlab.cern.ch/atlas/athena/-/merge_requests/22694Fix thread-unsafety in EscapedEnergyRegistry class - CaloCalibrationHits mt v...2019-05-08T03:02:59+02:00Marilena BandieramonteFix thread-unsafety in EscapedEnergyRegistry class - CaloCalibrationHits mt validation [ATLASSIM-4106]EscapedEnergyRegistry class was implemented as a singleton, so it was not thread-safe.
This affected the HITs of LArCalibrationHitDeadMaterial when running in MT with the CaloCalibration on.
Added the #ifdef G4MULTITHREADED directive to ...EscapedEnergyRegistry class was implemented as a singleton, so it was not thread-safe.
This affected the HITs of LArCalibrationHitDeadMaterial when running in MT with the CaloCalibration on.
Added the #ifdef G4MULTITHREADED directive to handle
the multithreaded case. One instance of the class will be created
per each thread and stored in a tbb::concurrent_unordered_map that
is hashed with the threadID number.
This change is implemented in CaloG4Sim package but affects
LArG4H6SD, LArG4H8SD, LArG4SD, TileGeoG4Calib packages, that are using EscapedEnergyRegistry class.
This was tested in sequential and mt mode with 20 ttbar events, and now the differences in the
LArCalibrationHitDeadMaterial are gone. Results between sequential and mt runs are consistent.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23224CaloRec: Avoid EventContext TLS lookups.2019-05-08T03:02:09+02:00Scott SnyderCaloRec: Avoid EventContext TLS lookups.Don't fetch EventContext from thread-local storage if we already have
it as an argument.Don't fetch EventContext from thread-local storage if we already have
it as an argument.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23074Adding new Exclusive Jet L1Topo algorithm2019-05-15T14:45:54+02:00Carlos Moreno MartinezAdding new Exclusive Jet L1Topo algorithmFollowing https://its.cern.ch/jira/browse/ATR-17320 a new L1Topo algorithm is implemented for exclusive jets. The cut values (xi_min, xi_max) are left free to be changed directly from the menu for more flexibility. They could be set to f...Following https://its.cern.ch/jira/browse/ATR-17320 a new L1Topo algorithm is implemented for exclusive jets. The cut values (xi_min, xi_max) are left free to be changed directly from the menu for more flexibility. They could be set to fixed values inside the algorithm but changing them afterwards would need a change inside the algorithm.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23219SGTools: Remove more forwarding headers.2019-05-08T03:02:03+02:00Scott SnyderSGTools: Remove more forwarding headers.Remove the obsolete forwarding headers CLIDRegistry.h, ClassName.h, and
ClassID_traits.h. They're no longer used from anywhere.Remove the obsolete forwarding headers CLIDRegistry.h, ClassName.h, and
ClassID_traits.h. They're no longer used from anywhere.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23225CaloUtils: Avoid EventContext TLS lookups.2019-05-08T03:02:52+02:00Scott SnyderCaloUtils: Avoid EventContext TLS lookups.Don't fetch EventContext from thread-local storage if we already have
it as an argument.Don't fetch EventContext from thread-local storage if we already have
it as an argument.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23207fix runTrigART for duplicate file names in PATH2019-05-08T03:02:28+02:00Rafal Bielskirafal.bielski@cern.chfix runTrigART for duplicate file names in PATHFixing an issue reported in ATR-19434.
The new script runTrigART.py was failing when one checked out TrigUpgradeTest because the same test was found in PATH from the release and from local build. This fix should take the first instance ...Fixing an issue reported in ATR-19434.
The new script runTrigART.py was failing when one checked out TrigUpgradeTest because the same test was found in PATH from the release and from local build. This fix should take the first instance found in PATH, which should be the local-build version.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23220CaloClusterCorrection: Avoid EventContext TLS lookups.2019-05-08T03:02:39+02:00Scott SnyderCaloClusterCorrection: Avoid EventContext TLS lookups.Don't fetch EventContext from thread-local storage if we already have
it as an argument.Don't fetch EventContext from thread-local storage if we already have
it as an argument.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23156Make SiCombinatorialTrackFinder_xk private (ATLASRECTS-3037)2019-05-17T01:11:57+02:00Susumu OdaMake SiCombinatorialTrackFinder_xk private (ATLASRECTS-3037)Make SiCombinatorialTrackFinder_xk private.
RunTier0Tests.py was fine. [RunTier0Tests.log](/uploads/731b4fb182985ef8fba3073518345c35/RunTier0Tests.log)Make SiCombinatorialTrackFinder_xk private.
RunTier0Tests.py was fine. [RunTier0Tests.log](/uploads/731b4fb182985ef8fba3073518345c35/RunTier0Tests.log)https://gitlab.cern.ch/atlas/athena/-/merge_requests/23205Only check whether the summary tool exists rather than retrieving it accident...2019-05-08T03:02:38+02:00Goetz GayckenOnly check whether the summary tool exists rather than retrieving it accidentially.In some cases the summary tool was retrieved by the validity check despite being disabled.In some cases the summary tool was retrieved by the validity check despite being disabled.https://gitlab.cern.ch/atlas/athena/-/merge_requests/2313621.2 egamma calib scales20182019-05-08T03:03:11+02:00Nansi Andari21.2 egamma calib scales2018New es model in the calibration tool to include 2018 scales, and update to R21 systematics.New es model in the calibration tool to include 2018 scales, and update to R21 systematics.Nansi AndariNansi Andarihttps://gitlab.cern.ch/atlas/athena/-/merge_requests/23190AnalysisTop: Updating TopBoostedTaggingCPTools to new recommendations2019-05-09T16:44:21+02:00Petr JackaAnalysisTop: Updating TopBoostedTaggingCPTools to new recommendations* Updated TopBoostedTaggingCPTools to the latest recommendations
* Improved flexibility:
* User can choose an arbitrary set of available taggers from the top config file
* `LargeRToptaggingConfigFile` option is replaced by `...* Updated TopBoostedTaggingCPTools to the latest recommendations
* Improved flexibility:
* User can choose an arbitrary set of available taggers from the top config file
* `LargeRToptaggingConfigFile` option is replaced by `BoostedJetTagging` option
* Usage:
* `BoostedJetTagging <TAGGER_TYPE>:<Tagger short name>`
* <TAGGER_TYPE> corresponds to TAGGER_TYPE in the BoostedJetTaggingRecommendationFullRun2 twiki. <Tagger short name> is the same as it is used in the boosted jet tagging tools to decorate large-R jets.
* Multiple taggers definition (works also w/o commas): `BoostedJetTagging JSSWTopTaggerDNN:DNNTaggerTopQuarkInclusive80, JSSWTopTaggerDNN:DNNTaggerTopQuarkContained50, SmoothedWZTagger:SmoothWContained50`
* Output variable: `vector<char> ljet_isTagged_<TAGGER_TYPE>_<Tagger short name>`
* validation-cuts.txt and ART tests config files are updated to reflect this change
See ANALYSISTO-793 for more details.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23203Fix.fast.pool.merge2019-05-08T03:02:31+02:00Marcin NowakFix.fast.pool.mergeFix bugs in fast POOL merging (both merging and reading)
Cleanup parts of the APR API for reading at the same timeFix bugs in fast POOL merging (both merging and reading)
Cleanup parts of the APR API for reading at the same timehttps://gitlab.cern.ch/atlas/athena/-/merge_requests/23212Make PanTau tools private (except those that have mutilple clients)2019-05-08T03:02:27+02:00Walter LamplMake PanTau tools private (except those that have mutilple clients)Make most of the tools used by PanTauProcessor private. Exception is the 'PanTau::Tool_InformationStore' that gets used by many clients. AFAICS, this is 'const-after-init' tool.Make most of the tools used by PanTauProcessor private. Exception is the 'PanTau::Tool_InformationStore' that gets used by many clients. AFAICS, this is 'const-after-init' tool.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23229Remove the no longer needed separate SpectrometerExtrapolator2019-05-09T03:02:28+02:00Pascal BoeschotenRemove the no longer needed separate SpectrometerExtrapolatorAs discussed in !22798:
MaterialAllocator doesn't need a separate SpectrometerExtrapolator anymore, its usage is now handled by the class's other extrapolator.As discussed in !22798:
MaterialAllocator doesn't need a separate SpectrometerExtrapolator anymore, its usage is now handled by the class's other extrapolator.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23141Propagate stream decisions from input to output2019-05-09T03:02:19+02:00Jack CranshawPropagate stream decisions from input to outputCheck for stream decisions in the input attribute list and
copy them to the output attribute list.
- Controlled through OutputStreamAthenaPool.py
- Default=OFF
- Two new tests in AthenaPoolMultiTestCheck for stream decisions in the input attribute list and
copy them to the output attribute list.
- Controlled through OutputStreamAthenaPool.py
- Default=OFF
- Two new tests in AthenaPoolMultiTesthttps://gitlab.cern.ch/atlas/athena/-/merge_requests/23227Sweeping !22608 from 21.0 to master.
Check against corrupted secondPCA file (...2019-05-09T03:02:22+02:00Atlas NightlybuildSweeping !22608 from 21.0 to master.
Check against corrupted secondPCA file (ATLASSIM-4141)Check against corrupted secondPCA file (ATLASSIM-4141)
See merge request atlas/athena!22608Check against corrupted secondPCA file (ATLASSIM-4141)
See merge request atlas/athena!22608https://gitlab.cern.ch/atlas/athena/-/merge_requests/22911Add empty and common steps in HLT CF menu2019-05-14T18:59:19+02:00Francesca PastoreAdd empty and common steps in HLT CF menu[HLTAllSteps.dot.pdf](/uploads/12e1240240dacc894d540dc57080b8dc/HLTAllSteps.dot.pdf)
Main changes in the CF builder to manage chains that share only parts of sequences. If two chains have a common step, but rest is different, the CF s...[HLTAllSteps.dot.pdf](/uploads/12e1240240dacc894d540dc57080b8dc/HLTAllSteps.dot.pdf)
Main changes in the CF builder to manage chains that share only parts of sequences. If two chains have a common step, but rest is different, the CF sequence is duplicated because one needs a new Filter+new Hypo (duplicated as well). This makes the graph more complicated to debug. See attached diagram.
Some changes in the CF builder to include the empty steps. Empty steps are managed as steps with one Filter Alg only, so successive step receives the output of the FilterAlg, instead of the Hypoalg.
Some changes are needed in the Dot functions, to envisage such cases.
The emu_step test has been made more modular, to allow preparing more complex and large menus in the unit tests.
Added more use-cases in this test.
Covering use cases in ATR-19600.
Many ref files are updated, need to ask the experts if reasonable results are reached with these changes.
https://gitlab.cern.ch/atlas/athena/-/merge_requests/23249TrigUpgradeTest: Avoid EventContext TLS lookups.2019-05-09T03:02:12+02:00Scott SnyderTrigUpgradeTest: Avoid EventContext TLS lookups.Don't fetch EventContext from thread-local storage if we already have
it as an argument.Don't fetch EventContext from thread-local storage if we already have
it as an argument.https://gitlab.cern.ch/atlas/athena/-/merge_requests/23052Variable bins in AthenaMonitoring2019-05-09T03:02:11+02:00Charles BurtonVariable bins in AthenaMonitoringThis MR is to handle the request to support variable bin-widths that was made in https://its.cern.ch/jira/browse/ATLASDQ-684. As a first step, I've done this only for TH1 histograms. If these changes are acceptable, I'll use the same met...This MR is to handle the request to support variable bin-widths that was made in https://its.cern.ch/jira/browse/ATLASDQ-684. As a first step, I've done this only for TH1 histograms. If these changes are acceptable, I'll use the same method for TH2, TProfile, and TEfficiency in another commit.
I've implemented this by allowing the user to specify a list of bin edges in the xbins kwarg of `GenericMonitoringTool.py`'s `defineHistogram` function. Then, in the `resolveAxes()` function of `HistogramDef`, they are parsed from a string into a new member variable `xArray`. This array is then passed to the appropriate `create<H,T>()` function of `HistogramFactory`.
@tbold @pisarna @ponyisihttps://gitlab.cern.ch/atlas/athena/-/merge_requests/23231Fix of FTK setting in tau trigger menu2019-05-09T03:02:07+02:00Bertrand Martin Dit LatourFix of FTK setting in tau trigger menuHello,
This MR is a fix discussed in ATR-19738.
We need to drop an FTK configuration setting that no longer exists.
Cheers,
BertrandHello,
This MR is a fix discussed in ATR-19738.
We need to drop an FTK configuration setting that no longer exists.
Cheers,
Bertrand