Astrophysics on GPUs for InterdiscipLinary Exascience challenges
Numerical codes solving fluid dynamical equations are a cornerstone of modern astrophysical research. Sophisticated codes with adaptive mesh refinement techniques are now routinely employed by the community, both to investigate fundamental physical processes and to model observational data. Our hunger for higher resolution and larger scale 3D simulations is rising and additional physics needs to be incorporated into the simulations. To deliver on these demands in a world of limited resources, our simulation techniques need to exploit computing hardware in the most efficient way.
In this project we will upgrade our open source community codes (AMRVAC framework: amrvac.org, bhac.science) to benefit from recent developments in HPC which invests heavily in accelerators (e.g. GPUs). The added performance will enable us to tackle challenging problems in solar- and high energy astrophysics. This effort is strongly supported by a healthy international user- and developer community.
MPI-AMRVAC is a parallel adaptive mesh refinement framework aimed at solving (primarily hyperbolic) partial differential equations by a number of different numerical schemes. The emphasis is on (near) conservation laws and on shock-dominated problems in particular.