Various Approaches to supercomputer modeling of multi-scale astrophysics challenges
Igor Kulikov
It is known that simulation of an individual filament with a size of several tens of megaparsecs with a scale of a normal galaxy size of ten kiloparsecs that require a computational grid of the order of 10K^3 is available nowadays [1]. However, the gap between the detailed description of galaxies and large-scale structures in the universe is unlikely to be solved in the near future. At the Mind the Gap 2013 conference (Cambridge, UK), Professor Volker Springel noted in his report that it would take about 60 years to achieve a resolution of the order of a single star in cosmological modeling while maintaining the extensive growth of computing capacity. However, in the next 12 years, exascale performance was achieved, albeit with a five-year delay, and zetascale computing is apparently expected closer to 2040. The problem is complicated by the lack of codes suitable for exascale supercomputers. In 2013, at the “Exaflos Computing in Astrophysics” conference (Ascona, Switzerland), challenges and a roadmap for developing core ExaScale codes were formulated. However, these issues remain unresolved, making ExaScale Astrophysics conferences relevant even in 2024. The main challenge is the parallel programming technologies. In recent years, graphic [2] and other accelerators [3], as well as vectorization of computations [4], have been actively used. However, using these technologies during code development, and especially in maintaining and extending the code, poses significant difficulties. Coarray Fortran technology [5] has become an alternative to traditional parallel programming tools, allowing for the development of scalable parallel code with complex architecture in a relatively short time. In addition to extensive numerical modeling development associated mainly with parallel programming, it is crucial to foster the intensive growth of computational codes focusing on physics-based multi-scale solutions. The “Mind the Gap” problem is not limited to cosmological modeling but also applies to modeling processes like star formation. Solutions include special computational grids [5] and mechanisms of detonation in white dwarfs, as well as subgrid functions for turbulent combustion development in white dwarf material [6] and carbon combustion scenarios during SNeIa explosions. Various approaches to tackling computational aspects of the “Mind the Gap” problem will be discussed in the report.
Event Timeslots (1)
Friday
-