Skip to content

Update BnoC isolation options HLT2

Francesca Swystun requested to merge bnoc_fswystun_isolation_update into bnoc_run3

MR to update the ioslation options for the BnoC WG.

This currently affects isolation options by:

  • Widening cone angle used to persist isolation variables
  • Removes unused code for vertex-based isolation

Cone isolation

As presented in the BnoC contribution at Analysis and Software week https://indico.cern.ch/event/1396208/#24-perspective-from-b2ccbnoc, we would like to increase the cone angle persisted at HLT2 level. This is increased from here from a persistency of dR^2<0.5 (allowing cone angle of deltaR up to ~0.7 offline) to dR^2<1.0 (deltaR 1.0 offline).

This is to account for the the cone angle typically used in the equivalent Run1/2 BnoC analyses varying from 1.3-1.7 whilst still being bandwidth conservative.

Updated bandwidth estimate using up-to-date bw-snippet. Tested on 50k minbias, expected_2024_min_bias_hlt1_filtered_v2. Use compression on output mdf file:

  • dr^2<0.5: 0.084 GB/s
  • dr^2<1.0: 0.086 GB/s (change proposed here)

Vertex isolation

Offline studies using RD, DaVinci options and persistency indicate the vertex-based isolation variables are not behaving as expected for BnoC lines.

Further study is ongoing to investigate this, for now remove outdated persistency options to prevent use in the WG.

cc (all those working on BnoC/B2CC isolation options) @oozcelik @yiming @dlancier @mmonk @tlatham @malehura @ghallett @masmith

Edited by Francesca Swystun

Merge request reports