ATLAS Software
Last updated
Last updated
If you already have some familiarity with ATLAS Software, you can probably skip this prologue style chapter (but it's not that long of a read!)
I'll briefly give an introduction to ATLAS Software and the Data Format.
For our purposes ATLAS software (SW) can be broken in two: Athena and Analysis Releases (it gets more complicated but that's beyond the scope of this small tutorial).
Athena: The main offline software - can do everything (simulation, reconstruction, analyisis, ...)
Analysis Releases: A subset of the main ATLAS code base (that exists in Athena) but can be used outside of Athena ("dual use code")... and also some analysis specific code on top of it.
All of the code is stored in a git repository hosted at the Athena GitLab page .
People who run the core software group and the analysis software group handle building what parts of the code are needed for Athena releases and what parts are needed for analysis releases. The builds are called releases and they are installed to the ATLAS CVMFS areas so that all machines connected to the ATLAS CVMFS repositories have access to the releases. These days, all code is built with cmake. We'll talk more about that later.
The current main release of ATLAS Software is Athena release 21.0.X.
The Analysis Releases are on the numbering scheme 21.2.X. For the main part of this workbook, we'll be setting up and using an Analysis Release.
The data format (what is actually stored on disk in ROOT files) is very closely related to the analysis software. The xAOD model was developed for Run 2 of ATLAS. The xAOD model is a set of classes which make it nice and intuitive to analyze physics objects. At it's core it represents physics objects as C++ classes (e.g. Event, TrackParticle, Electron, Jet, etc.). It also provides the means to analyze those objects, store them in ROOT files, and "decorate" them with properties.
The actual files stored on disks all over the world at grid sites and in your work areas if you've downloaded some from rucio are called DAOD's (Derived Analysis Object Data, some people also call the files themselves xAOD's). A primary AOD is the starting point: Some Athena code produces the AOD after reconstruction, and then the Derivation Framework is used to produce the Derived AOD's. These are tailored for specific purposes. Starting from the AOD which contains everything, a derivation is made by removing whole events, removing whole objects from events, removing some information attached to objects in some events, adding some information to some objects (we do this in the TRT a lot!), or calibrating existing objects. The nomenclature:
skimming: removing whole events
thinning: removing whole objects from within an event, but keeping the rest of the event
slimming: removing information from within objects, but keeping the rest of the object
augmentation: adding data not found in the input data
In the TRT SW group, we use the Inner Detector (InDet) specific derivation, with some of our own stuff built on top.
That's all I'm going to say about ATLAS SW here since we need to get onto TRT specific things -- but there's a lot more detailed information in other places and I'd suggest starting !