Simulations enable us to investigate various earthquake scenarios and provide a deeper understanding of the real-world earthquakes never happening twice. That is why it is profitable in elucidating the earthquake source process.
Having said that, real-problem solvability is often restricted by our computational resources. Here we tackle developing a fast solver for the earthquake dynamic rupture problems.
Sato & Ando (2021) developed a fast algorithm for the space-time domain boundary element (boundary integral equation) method applicable to any fault geometry. The computation time per step and the total memory usage are approximately of order the number of elements. It is currently the smallest order, besides seemingly lighter than all previous versatile ones wrt the memory.
The paper is for 2D as it treats the algorithm development. The next will be the application to 3D.