Tools and studies for fast simulation of calorimeter showers: detailed Geant4 / DD4hep simulation, compact point-cloud representations of the showers, and their use as input to machine-learning-based fast simulation. Some trained weights are on Hugging Face.
One photon shower as Geant4 steps and after two step2point compressions
(made with render_shower_display.py).
Most recent first, by arXiv date.
| Paper | Journal | Code | Data and weights |
|---|---|---|---|
| step2point dataset: Detailed shower simulation for data representation studies, arXiv:2509.22340 (CDS) | step2point |
Zenodo, doi:10.5281/zenodo.17199427 |
| Dataset | Content | Where |
|---|---|---|
| step2point dataset (2025) | Detailed (Geant4 step-level) simulation of electromagnetic and hadronic showers in the Open Data Detector calorimeters, for studies of shower data representations | Zenodo, doi:10.5281/zenodo.17199427 |
| Repository | Content |
|---|---|
step2point |
Library for compressing calorimeter shower data (x, y, z, E) into point clouds while preserving key physics observables. Documentation |
ddFastSim |
Plugins for fast simulation with DDG4, starting with a replication of the Geant4 extended/parameterisations/Par04 example |
DDsimTutorialApplication |
Reproduction of the thin-target Geant4 tutorial application with DD4hep simulation (DDSim) |
showerRepresentationStudies |
Tools to analyse shower data for data-representation optimisation in ML-based calorimeter simulation (CERN-HSF project) |
To add a paper, a dataset or a repository, open a pull request editing profile/README.md in
fast-sim/.github.