Skip to content
GitLab
v30r0 protected
Marco Cattaneo's avatar
Marco Cattaneo
@cattanem
26a7dae3 · Merge branch 'v30r0-release' into 'master' ·
Release v30r0

2018-03-22 Rec v30r0

This version uses Lbcom v30r0, LHCb v50r0, Gaudi v30r2 and LCG_93 with ROOT 6.12.06

This version is released on `master` branch. The previous release on `master` branch was `Rec v22r1`.

This release includes the merge to master of the future branch, done in November 2017 - see !800 commit history for details of future changes that were included

This release also contains all the changes that were made in Rec v23r0 (released on 2018-patches branch, see corresponding release notes), as well as the additional changes described in this file.

New features

Enhancements

  • Template TrackListMerger and export copies for LHCb::Tracks and LHCb::Track::Selection, !958 (@olupton)

  • Faster PrVeloUT, !859, !913 (@decianm)

    • Get rid of the tool, do everything directly in PrVeloUT
    • Cache look-up table for tolerances locally, instead of accessing the tool for every track. Still need to do the same for the Bdl table.
    • Avoid using a std::vector in the clustering stage, as it just creates unnecessary overhead
    • Introduce a minimum momentum cut of 1500MeV. More detailed checks needed for possible efficiency loss
    • Speed up hit search by precalculating some quantities.
    • Don't split hits in top and bottom anymore, gain was not worth the savings in time.
    • Also contains a (prototype) version of the method that creates the output tracks, as most of the time is spent in the method that writes out the LHCb::Track in keyed containers. In principle only x,y,z,tx,ty,q/p and the LHCbIDs are needed from VeloUT

  • Try to fix FT zone order, !891 (@mhadji)

  • Improvements for faster decoding in PrFTStoreHit. Reduces CPU consumption by ~40%, !882, !897 (@jvantilb) [LHCBSCIFI-109]

  • replace isLarge with pseudoSize for FTLiteClusters, resolution depends on pseudoSize, !838 (@sesen)

  • Adapt to new templated LHCb traj for TVF, !894, !914 (@mhadji)

  • declare MeasProviders as properties, !898 (@sstemmle)

  • Use addSortedForwardStates when inserting new states on TrackEventFitter, !847 (@mhadji)

  • Vectorized updateTransport in TrackVectorFitter, !811 (@dcampora)
    Closes #2

  • Removed reference nodes from TrackVectorFitter, !802 (@dcampora)
    Closes #4, #6

  • changes in MuonID after 12/2017 hackathon, !922 (@rvazquez)

    • change vdt::fast_exp to std::exp as profiling shows it to be faster
      • remove the Choleskiy decomposition for matrix inversion in favour of Kramer as profiling shows it to be faster
      • change when possible double to float
      • change when possible int to unsigned int

  • Fixes for running muon algorithms on upgrade MC, !823 (@olupton)

  • MuonID upgrade to Gaudi functional, !822 (@nkazeev)

  • Rich adapt to decoded data format update, !923 (@jonrob)

  • Various RICH CPU improvements, !907 (@jonrob)

    • Adapt to move of SIMD mirror data cache to RichFutureUtils in LHCb.
    • Use new SIMD global to PD panel coordinate system transformation in the smart ID tool.
    • Small improvement to the detectable signal algorithm to avoid unnecessary find calls.
    • CPU improvements to the geometrical efficiency algorithm from converting the TES data type from map<A,B> to vector< pair<A,B> >.

  • RichRayTracing - Turn off the beam pipe intersection checks with raytracing photons for the CK rings., !889 (@jonrob)

  • RICH - Turn off by default unambiguous photon and beam pipe intersection checks in quartic reconstruction., !888 (@jonrob)

  • Reduce secondary mirror iterations to 1 for RICH1 in quartic reconstruction, !886 (@jonrob)

  • Reoptimise the RICH photon selection cuts for the Upgrade, !875 (@jonrob)

  • RICH - Minor optimisation, !848 (@jonrob)

    • Add to reco base class methods to retrieve list of active mass hypotheses excluding the below threshold type.
    • Remove unnecessary normalisation of final photon direction in quartic reconstruction.
    • Remove some residual aerogel support code.

  • Small update to RICH expected photon signal algorithm, !830 (@jonrob)

    • Remove the Below Threshold type from the hypo list during initialize() rather than skip each time.
    • Bail out the hypo loop as soon as a below threshold type is found.
    • Make the min expected track CK theta angle a configurable radiator property.
    • Drive by cleanup of RichFutureRecBase.

  • Convert RICH CK Estimation photon reconstruction alg to SIMD version., !831 (@jonrob)

  • Use fully SIMD RICH ray tracing API, !795 (@jonrob)

  • Fully SIMD vectorised RICH photon reconstruction, !779, !814, !884 (@jonrob)

  • Backported optimizations in memory allocations from TDR to master, !860 (@sponce)

  • Adapt to Future linker base class, !839 (@sstahl)

Bug fixes

  • Suppress a maybe unitialised variable warning in StandaloneMuonRec, !971 (@jonrob)

  • Remove default muon state, !966 (@sstahl)
    The state is not produced by the MuonID, only by some monitoring.

  • Remove duplicate MuonIDAlgLite properties, !955 (@olupton)

  • PrVeloUT: Fix sometimes off by one in degrees of freedom, !947 (@graven)

  • Fix typo in PatLongLivedTracking parameters and fix some indentations, !941 (@decianm)

  • Fix in-place modification of PrPixelTracking property, !927 (@olupton)

  • Fix undefined reference with -O0 after gaudi/Gaudi!514, !903 (@clemenci)

  • Fix TrackListRefiner for VPTrackSelector in Alignment, !825 (@cburr)

  • Fix clang warnings, !819 (@cattanem)

    • ParameterizedKalman: replace abs by std::abs
    • CaloMoniDst: remove unused variables

  • Fix typo in assertions (std::none_of -> std::all_of), !812 (@chasse)

  • Fix typo in PrMatchNN, !806 (@chasse)

  • Fix trivial clang warnings, !807 (@cattanem)

  • Fix clang compilation warnings, !784 (@cattanem)

    • PrForwardTool: remove unused member variable
    • TrackSTEPExtrapolator: Remove unused member variables

  • Fix clang warnings in MuonID, !909 (@cattanem)

  • Prevent automatic retrieval of undesired ToolHandles, !791 (@clemenci)

Code modernisations and cleanups

  • Generalize IMeasurementProvider: prefer span over iterator_range, !940 (@graven)

  • PrVeloUT: implicitly deduce template argument, !943 (@graven)

  • Added method in PrUTMagnetTool/PrTableForFunction to return the table, !855 (@decianm)

  • Do not assume that Track::states() returns a reference to a container, !944 (@graven)

  • Removed Tr::TrackVectorFit::FitTrack, !753 (@dcampora)

  • Adapt TrackBestTrackCreator to automatic ToolHandle retrieval in gaudi/Gaudi!429, !782 (@cattanem)

  • Generalize TrackTraj, !915 (@graven)

    • avoid mentioning StateContainer, accept span<State* const> instead
    • prefer offsets over iterators
    • use assert instead of commented out checks
    • make static object const

  • Modernize TrackFitter, !849 (@graven)

    • Fully qualify enum values
    • Prefer explicit enums over ints
    • Prefer control flow which reduces indentation

  • Modernize Track Likelihood, !837 (@graven)

    • avoid redundant copies & loops
    • add ITrackFunctor in order to be able to deprecate the use of ITrackManipulator
    • Make TrackLikelihood inherit from ITrackFunctor instead of ITrackManipulator, and make it const-correct
    • TrackAddLikelihood now calls ITrackFunctor to obtain values, and then it forwards those value to Track::setLikelihood
    • Prefer Gaudi::Property

  • Generalize ITrackVertexer interface by using span instead of an explicit container type, !917 (@graven)
    and update the TrackVertexer implementation accordingly

  • Modernize TrackCheckerNT, !931 (@graven)

    • avoid redundant casts
    • prefer STL algorithms over raw loops
    • prefer auto
    • prefer switch/case over if-else-if-else cascades
    • avoid explicit memory allocation

  • Further modernize TrackCaloMatchMonitor, !932 (@graven)

    • prefer Gaudi::Property
    • inherit constructor
    • do not use front and back of ConstNodeRange (to allow ConstNodeRange to become a gsl::span)
    • use pointer / boost::optional for 'nullable' State instead of an explicit bool flag
    • prefer emplace_back
    • prefer std::accumulate over explicit raw loop

  • Modernize TrackMonitors, !916 (@graven)

    • Avoid mentioning LHCb::Track::ConstNodeRange
    • prefer switch over if-else-if-else-
    • prefer TupleTool::columns over TupleTool::column

  • RichRecUtils - Remove unused code, !854 (@jonrob)

  • RICH Clean up Pixel SIMD algorithm, !828 (@jonrob)

  • Remove old RICH framework + Future scalar algorithms, !818 (@jonrob)

  • Remove unused captures in Rich global pid algorithm, !778 (@jonrob)

  • Modernize Muons, !717 (@graven)

    • make interfaces inherit from extend_interfaces, and use DeclareInterfaceID
    • avoid trailing return types in interfaces
    • deprecate interface methods which take a string as argument, provide method which takes an enum instead -- this avoids having to create a string, and then parse it, on every call
    • prefer inheriting constructors
    • prefer std::abs over fabs
    • prefer std::transform over std::for_each
    • add const
    • move debug printout out of loop
    • simplify logic
    • prefer std::array over std::vector to avoid heap allocation
    • prefer passing std::array<>& over decayed pointers
    • prefer true/false over 1/0 assignments for bool
    • prefer std::array.fill over loops
    • simplify inside using lambda
    • prefer range-based for loops
    • prefer switch/case over if-else-if-else-if-
    • prefer boost::static_vector over std::vector if there is a known upperlimit on the maximum size
    • prefer range-based loop over BOOST_FOREACH

  • msgLevel already caches the level -- do not cache the cached value, !890 (@graven)

  • Fully qualify enum values, !851, !852, !853, !856, !868 (@graven)

  • Adapt code to LHCb!1077, !865 (@graven)

    • avoid using eventSvc() in MCTrackInfo, prefer make_MCTrackInfo
    • do not explicitly fetch MCParticles from event store
    • remove explicit initialize and finalize
    • prefer Gaudi::Property

  • Remove extraneous PUBLIC_HEADERS uncovered by gaudi/Gaudi!615, !933 (@rmatev)

  • Removed use of deprecated factories, !803, !964 (@clemenci)

    • replaced DECLARE_*_FACTORY with DECLARE_COMPONENT
    • made constructors public
    • removed some unnecessary disabling of copy constructors
    • removed some trivial destructors
      See gaudi/Gaudi!420

  • Adapted to change in logical comparison between StatusCodes, !885 (@clemenci)
    see gaudi/Gaudi!514

  • Forward compatibility for gaudi/Gaudi!514, !870 (@clemenci)

  • Make forward/backward compatible with gaudi/Gaudi!534, !805 (@graven)

Monitoring changes

Changes to tests