Introduction
The Geant4 simulation framework which we call “dgcode” is implemented as a set of simplebuild packages (specifically in the dgcode bundle). Its history goes back to 2012, but in 2023/2024 it was cleaned up and resurrected in its present form.
Most importantly, dgcode includes functionality for easy setting up and launching of Geant4 simulation jobs based on configurable geometry and particle generator modules, and contains various functionality for material definitions, physics lists choice, visualisation, multiprocessing, random number handling, output persistification and analysis, etc.
Before starting to use dgcode, you should first of all make sure you have installed it, and secondly you should spend some time with the simplebuild documentation at https://mctools.github.io/simplebuild, since working with dgcode means working within the simplebuild build system. It will eventually also be a good idea to dive into parts of the Geant4 documentation, in particular the parts about geometry definitions (note that for historical reasons geometry definition is often referred to as “Detector definition” in Geant4).
Features
Geometry is written in C++ as separate modules.
Particle generators are likewise written as separate modules with tunable parameters, but may also be written in C++. It is also often the case that a pre-existing generator can be used.
Overall simulation control is configured in a small Python script known as a sim-script, in which one will:
Choose and configure geometry and particle generator modules.
Optionally change other defaults of the script, including tweaking the simulation physics and configure output data as needed.
The sim-script is then essentially a full-featured application which can be launched from the command line. It provides a lot of functionality which can all be invoked by adding flags at the command line. Examples include:
Specify number of events, initial random seed and number of processes in case multi-processing is desired.
Select physics list (either reference G4 lists or custom lists which can also be written as optional modules).
NCrystal is always enabled, ensuring high fidelity scattering cross sections for thermal neutrons.
Launch 3D viewers or interactive Geant4 sessions.
Control output in Griff or MCPL format, or collect custom heat maps.
Allow tuning of geometry and particle generator parameters, which is useful both for parameter scanning and for simply “playing around” and trying to understand a simulation.
Post-simulation analysis can be done as desired, and might for instance include a Griff analysis written in C++, which outputs histogram data in SimpleHists files, usually followed by a final statistical analysis and plot-production in Python, where all the usual tools (e.g. Matplotlib and SciPy) are available.
Due in particular to being based on simplebuild, dgcode readily facilitates the creation of shared work between multiple related Geant4 projects within a group of people.
How to use the documentation
New users will most likely benefit from first following the installation instructions, to ensure that dgcode and simplebuild are available on the system. It is probably best to then spend a bit of time on the simplebuild documentation, focusing at least on the introduction and usage examples. Then, it is probably time to try to follow the instructions for how to easily start a dgcode-based simulation project.
After these initial steps, it will be time to start diving into the more detailed parts of the documentation, which you can find either from the sidebar menu or via the subjects overview page. You can of course also use the documentation’s search functionality in case you can not easily find something.