Native Julia HEP Geometry package

Description

The Julia programming language offers a unique combination of speed, interoperability, ease of use, and flexibility, making it an attractive option for High Energy Physics (HEP) research. Within the HSF context, the JuliaHEP initiative aims to develop a set of foundational packages in the Julia ecosystem that provide the essential functionality required by HEP researchers. In simulation and reconstruction codes, it is necessary to describe the detailed geometry of the detector, including both its spatial structure and the associated materials, surfaces, and other properties. This is typically achieved using a hierarchical Constructive Solid Geometry (CSG) representation, as implemented in frameworks like Geant4. Reconstruction codes may, however, require a simplified version of this geometry.

The Geant4.jl wrapper package currently leverages the geometry modeling capabilities of the C++ Geant4 library. However, if we wish to develop physics process simulations natively in Julia and independent of Geant4, a geometry package implemented directly in Julia will be required. Such a package should aim to be user-friendly, high-performance, and compatible with hardware accelerators such as GPUs. An early effort in this direction is the Geom4hep package, which could serve as a potential foundation for further development. In addition, the CSG-system that is currently part of SolidStateDetectors.jl can be used as a baseline. In an initial application, the geometry can be used to export a high-quality mesh with adaptive level of details to allow raytracing-based optical physics simulation leveraging modern RTX GPU functionality. This would allow hybrid Monte-Carlo simulation infrastructures where Geant4.jl handles the ballistic particle-tracking while the optical simulations are handled by a custom julia code.

Task ideas

Design a standalone physics-compatible Julia CSG package with faithful import from and export to Geant4 via a package extension. The Geom4hep.jl package and the CSG-system that is currently part of SolidStateDetectors.jl can be used as a baseline.
Implement a high-quality mesh-export with adaptive level of detail, both for visualization (e.g. via Makie.jl) and for raytracing-based optical physics simulations.
Benchmark the developed implementation using raytracing simulations to evaluate performance and accuracy against existing solutions.
Implement a Geant4-like ballistic particle tracking that allows massive CPU/GPU parallelization by using an entity component system (ECS) like approach that groups particles by current volume type and by active physics. This would leverage Julia’s ability to generate GPU kernels at runtime, allowing kernel fusion of boolean geometry (sub-trees).

Expected results and milestones

Familiarization with existing HEP geometry packages in Julia and C++
Successful development of a new julia package with the basic geometry shapes
Implementation of geometry import/export functionality with Geant4
Implementation of mesh export with adaptive level of detail for visualization and raytracing
Development of a set of benchmarks for comparing the performance with respect other packages (e.g. Geant4.j)

Requirements

Programming experience with C++ (in order to be able to understand existing HEP codes)
Prior experience in Julia (very advantageous)
A background understanding of high-energy physics (advantageous)
Knowledge of geometry representations (CSG, meshes, etc.) in modern computer science and physics simulations (advantageous)
Experience with GPU programming (advantageous)

How to apply

Once CERN/HSF is accepted as a GSoC org, please write an email with a short introduction to your interests and background to the mentors with the string “gsoc26” in the subject. There will be a small evaluation task that we will mail to you then.

Mentors

Florian Henkes - TUM
Pere Mato - CERN
Oliver Schulz - MPP

Additional Information

Difficulty level (low / medium / high): medium
Duration: 350 hours
Mentor availability: June-October

Corresponding Project

JuliaHEP