Software development is a central and critical component of the Belle II experiment, in which the German groups are heavily involved and play a leading role in the (further) development of the Belle II experiment software. To ensure efficient and smooth operation of the experiment, components for automated creation, testing, documentation, and installation are essential. The necessary infrastructure for software development as well as central components of the software itself were largely created by German groups.
In addition, German groups classically play a strong role in track reconstruction. A major challenge in this area is again the underground, which will continue to intensify as the luminosity increases. To meet this challenge, background filters and calibration methods, among others, are constantly being developed and improved. In general, the resource requirements of track reconstruction are a major challenge as well, as despite increasing data volumes, the execution time must be kept within a narrow range to run on the high-level trigger. These adjustments require optimizations in many parts of the code and close collaboration between the German groups.
A special feature of the Belle II project is its high precision and efficiency in reconstructing particles, in particular neutral particles, photons and K⁰-mesons. These reconstructions require a particularly powerful and robust discrimination of particle types based on their signal shape. The algorithms required for this purpose are constantly being adapted, improved and automated in order to identify the particles more effectively. Modern technologies such as deep neural networks are employed in this process.
The development of analysis tools belongs to the field of software development as well. Many such analysis tools, which are used as standard tools in collaborations, originate from German groups. For example, the Full Event Interpretation (FEI), which can reconstruct one of the two B mesons in a Y(4S) event, is very successful. Other efforts include the development of tools to measure time-dependent CP asymmetries, the determination and automation of systematic uncertainties in the track reconstruction, the documentation of the software, and the training of the users who perform the physical analyses.
Another challenge for the Belle II experiment, though more related to computing hardware than software, is obtaining and managing storage capacity and computing power. Large computing resources are required to process the recorded data and generate simulation data. The Belle II collaboration has decided that resources should be provided by member countries through grid interfaces. German groups provide storage space in the peta-byte range on which raw data are stored, among other things, and are thus well visible as large Belle II centers. So far, this infrastructure has only been used for centrally coordinated tasks, such as simulating or reconstructing data - leaving the analyses to the individual institutes. However, since the data volumes for analyses will soon exceed the capacity limits of individual computers, German groups are working together with computing experts to develop an analysis infrastructure that will enable effective high-performance analysis with centralized distribution of resources.
Skill Box
Skill Box
To participate in the software development of Belle II, the following skills are required:
- Basic understanding of physical processes, detector technologies and analysis methods.
- Modern methods of software development, programming languages C++ and Python
- Ability to work in a team, good English skills and interest in working in globally distributed teams
