Magnetic materials are essential for modern applications, ranging from renewable energy sources to e-mobility. Improving their performance requires understanding magnetization reversals under different conditions. Magnetic domain structures, the building blocks of magnetic materials, interact in complex ways influenced by composition, microstructure, and external factors, such as applied magnetic fields. Experimentally, direct observation of magnetic domains can be performed at macroscopic length scales, providing information about their motion. However, physical understanding cannot be obtained so straightforwardly because the lateral resolution is limited by the wavelengths of visible light. Another way to investigate the magnetic properties of materials experimentally is to measure their bulk magnetization response, which provides an overview of how the material responds to external stimuli. While both experiments provide information about magnetic properties, a complete understanding is not possible without micromagnetic simulations. Micromagnetic simulations allow us to create detailed virtual representations of magnetic materials and study their behavior in different scenarios. However, such simulations are computationally demanding because they generally consist of a few million regions in three-dimensional space. Because these simulations are computationally demanding, parallel processing is necessary to simulate realistic structures that resemble real-world samples. Realistic simulation models require solutions to dynamic differential equations, making high-performance computers (HPCs) essential. Without HPCs, these calculations would take weeks to months of GPU time per simulation, making exploration of the parameter spaces impossible for a proper explanation of the experimental findings.