A Fast Simulation Suite for Cherenkov Detectors at the future Electron Ion Collider

Abstract

The integration of Deep Learning (DL) into experimental nuclear and particle physics has driven significant progress in simulation and reconstruction workflows. However, traditional simulation frameworks such as GEANT4 remain computationally intensive, especially for Cherenkov detectors, where simulating optical photon transport through complex geometries and reflective surfaces introduces a major bottleneck. To address this, we present an open, standalone fast simulation tool for Detection of Internally Reflected Cherenkov Light (DIRC) detectors, with a focus on the High-Performance DIRC (hpDIRC) at the future Electron-Ion Collider (EIC). Our framework incorporates a suite of generative models tailored to accelerate particle identification (PID) tasks by offering a scalable, GPU-accelerated alternative to full GEANT4-based simulations. Designed with accessibility in mind, our simulation package enables both DL researchers and physicists to efficiently generate high-fidelity large-scale datasets on demand, without relying on complex traditional simulation stacks. This flexibility supports the development and benchmarking of novel DL-driven PID methods. Moreover, this fast simulation pipeline represents a critical step toward enabling EIC-wide PID strategies that depend on virtually unlimited simulated samples, spanning the full acceptance of the hpDIRC.

Contents

• Data used for Training
• Environment
• Available Architectures
• Usage
• Reference Packages

Data used for Training

The data used for training our generative models was created using a standalone simulation. More details can be found at eicdirc. Our dataset was constructed over a singular bar (no azimuthal angle variance), and without magnetic field. In future updates, we will provide generative models trained under these scenarios.

Environment

Noteable requirements:

• Python: 3.12.8
• Pytorch: 2.5.1
• CUDA: 12.4
• Framework for Easily Invertible Architectures
• Flow Matching
• NFlows

The dependencies for the networks can be installed with the following command:

conda env create -f env.yml

In the case that some packages do not install through the provided conda command, you can install them using pip once your conda environment is activated:

python3 -m pip install <package>

Available Architectures

Our package currently consists of 5 state-of-the-art generative models, each providing fast simulation for both Pions and Kaons in the hpDIRC:

• Discrete Normalizing Flows - Default simulation method
• Continuous Normalizing Flows
• Flow Matching
• Score-Based Generative Models
• Denoising Diffusion Probabalistic Models (DDPM)

The default simulation method provides both the fastest, and highest quality generations.

Usage

Note we have provided all code required to reproduce the results found within the paper, although these require large amounts of simulation from GEANT4. For those interested in training their own models, or reproducing our work, please open an issue. If their exists a high demand, we will update our documentation to and provide instructions for dataset creation, model training and model evaluation. We also plan to release a streamlined, containerized version for those who only wish to utilize the fast simulation methods for dataset creation.

Running Fast Simulation

We have provided an example script to allow simulation at both fixed kinematics, and continuously over the phase space. An example command and argument details is provided below:

python run_simulation.py --config config/config.json --n_tracks {} --n_dump {} --method {} --momentum {} --theta {} --model_type {} --fine_grained_prior --dark_rate --dark_noise 22800

Argument	Type	Default	Description
--config	str	config.json	Path to the config file
--n_tracks	int	1e5	Number of particles to generate
--n_dump	int	None	Number of particles to dump per .pkl file
--method	str	"MixPK"	Generated particle type (Kaon, Pion, or MixPK)
--momentum	str	"3"	Momentum to generate (e.g., "6", "1-10")
--theta	str	"30"	Theta angle to generate (e.g., "30", "25-155")
--model_type	str	"NF"	Which model to use (NF,CNF,FlowMatching,Score,DDPM)
--fine_grained_prior	flag	True	Enable fine-grained prior (just include the flag to activate the option)
--dark_noise	flag	False	Whether or not to include dark rate in generations
--dark_rate	float	22800	Dark rate - default corresponds to -c 2031 in standalone Geant simulation
--use_gpu	flag	False	Whether to use a GPU - note CPU can be faster.
--num_threads	int	None	Number of CPU threads - default is pytorch defaults (1/2).

Given the way our models create individual tracks, we are actually capable of faster generations on CPU (i9-14900K is 2x as fast as RTX4090). Given relatively low model overhead and aggregation strategy, using a GPU can potentially lead to losses in performance due to data transfers. We reccomend testing this on your system. We also reccomend testing the number of threads argument, in which providing more threads may not increase performance. Generally the default pytorch values will suffice.

Hit Pattern Creation

For hit pattern creation we still require GPU in order to leverage large batch sizes of photons. In future updates we will also allow CPU generation, as some work is required to figure out scaling of memory overheads.

We have provided an example script to generate visualizations of hit patterns at fixed momentum (user provided), and over various theta in 5 degree bins. This will be done for both pions and kaons within the same script by default.

python plot_PDF.py --config config/config.json --method {} --momentum {} --model_type {} --fine_grained_prior --fs_support {}

Argument	Type	Default	Description
--config	str	config.json	Path to the config file
--momentum	float	6	Momentum to generate (e.g., 6.5, 1.25)
--model_type	str	"NF"	Which model to use (NF,CNF,FlowMatching,Score,DDPM)
--fine_grained_prior	flag	False	Enable fine-grained prior (just include the flag to activate the option)
--fs_support	float	False	Number of photons to generate per PDF (default 200k)

Reference Packages

The following is a list of references used in the creation of this repository.

• OT-Flow
• denoising-diffusion-pytorch
• FastDIRC
• Deep(er)RICH

Citation

@misc{giroux2025generativemodelsfastsimulation,
      title={Generative Models for Fast Simulation of Cherenkov Detectors at the Electron-Ion Collider}, 
      author={James Giroux and Michael Martinez and Cristiano Fanelli},
      year={2025},
      eprint={2504.19042},
      archivePrefix={arXiv},
      primaryClass={physics.ins-det},
      url={https://arxiv.org/abs/2504.19042}, 
}