Skip to content

Tips and tricks

Running old DaVinci on the GRID

The run 1 \(R(D^{(*)})\) analysis used DaVinci/v36r1p2. To run that on lxplus7 nodes:

  1. Export the following environmental variables to use old runtime:

    export CMTCONFIG=x86_64-slc6-gcc48-opt

  2. Run old DaVinci with lb-run:

    lb-run DaVinci/v36r1p2 gaudirun.py

About GaudiSequencer

GaudiSequencer is an per-event algorithm, such as DecayTreeTuple, in the sense that Gaudi knows that it should run on every event.

GaudiSequencer is used to chain related algorithms, and it's default behavior is to only run an algorithm if the preceding algorithm passed (this is done in the c++ class for the algorithm by returning StatusCode::SUCCESS).

This default can be overridden by setting GaudiSequencer.IgnoreFilterPassed to True, which will run all the algorithms regardless of whether the preceding one passed or failed. The list of algorithms to run is defined by the GaudiSequencer.Members list.

Difference between MainSequence and UserAlgorithms

  • MainSequence: Use DaVinci().addToMainSequence([<a list of algorithms>]) to customize the initialization of DaVinci. This should be used to provide custom functions/TESs1 that will be visible to all selection algorithms.

  • UserAlgorithms: Actual selection algorithms. This is always appended at the end of the MainSequence of a DaVinci session.

Difference CC and cc in a decay descriptor

  • LoKi and DecayTreeTuple use uppercase CC for charge conjugation.
  • CombineParticles uses lowercase cc.

This is just an inconsistency in the DaVinci code base and there's nothing we can do as a user.

  1. TES: Transient Event Storage. ↩