Campaign 2026-04
Unlike the other campaigns, this one is not a full production run. It is a focused investigation into two Lambda-related MCParticle bugs affecting the ePIC + DD4hep simulation chain.
Bugs under investigation
MCParticle endpoints / daughter tree wrong for Lambdaseic/epic#1069
Tracking-volume G4 handler drops Lambda decay infoeic/epic#1081 (review)
The symptom in both cases is the same: primary Lambdas in the output MC record either have no daughters, have wrong endpoint coordinates, or are replaced by a shower-like daughter list — which breaks any downstream analysis that follows the Λ -> p π- or Λ -> n π0 decay chain.
Variants produced
Only DD4hep + afterburner were re-run; no full reconstruction. The runs differ only in how the Geant4 tracking-volume handler and the --part.keepAllParticles flag are configured:
| Folder tag | What was changed |
|---|---|
saveall | --part.keepAllParticles=True (does not actually work due to DD4hep bug) |
notv2 | Tracking-volume G4 handler switched off -> true "save all" behaviour |
notw | Same as notv2 (handler off) |
longtv | Updated, longer tracking volume from the issue above |
Data location (ifarm)
/work/eic/users/romanov/meson-structure-2026-04-check/
afterburner/
dd4hep_2026-03/
dd4hep_26-04/
dd4hep_26-04_notv2/
dd4hep_26-04_saveall/
dd4hep_26-04_longtv/
dd4hep_26-04_notw/
dd4hep_26-03_saveall/
ana_26-04/
ana_26-04_notv2/
csv_dd4hep_26-04_notv2/Reports
- Tracking volume (saveall variant)
- Tracking volume comparison — saveall vs longtv (pending data regen)