ACTS meets BGV

Monday 8 Jun 2020, 14:00 → 15:00 Europe/Zurich

Present:

• Andreas Salzburger, Moritz Kiehn (EP-ADP)
• Bernadette, Helene, James (BE-BI)

Bernadette presented an overview of the BGV project, in particular, the simulation goals & workflow.

Discussion

BGV observations by Andreas & Moritz

• Forward type detector with planar sensors, very similar to a testbeam telescope
• Suggest to tilt detectors to improve vertex resolution, minimise detector material & improve alignment

Suggest to describe detector geometry using DD4HEP format:

• https://dd4hep.web.cern.ch/dd4hep/
• Advantage - provides a common detector description for simulation, reconstruction & analysis
• Use the same geometry description for both G4 & ACTS
• Two steps: 1) define layers in XML file and 2) write translator
• Thereafter translation into ACTS geometry very easy
• Enables truth fitting - rewind truth records & put in a filter

ACTS

• https://acts.web.cern.ch/ACTS/
• https://cds.cern.ch/record/2243297/files/ATL-SOFT-PROC-2017-030.pdf
• Toolkit, not a framework ( you call us, we don't call you! ), C++
• Based on the ATLAS tracking code - 20 years of development! 
• Development workflow
  • Docker containers ( use docker terminal for compilation, mount files locally for local editing ) - prefered option for local development
  • Lxplus - point to LCG releases

Misc. advice

• For parameter optimisation (scans, baysian, AI) keep everything in a fast pipeline
• Event loop (processing) framework
  • ACTS test framework
    • Config. (steering) layer
    • Event data handling ( write / read to whiteboards )
    • Event loop ( parallised in multiple threads )
  • In the long term GAUDI-Lite could be useful event loop framework
• Moritz developed specialised reconstruction package for beam telescope which could be useful (https://gitlab.cern.ch/proteus/proteus)