A multiblock compressible Navier-Stokes solver in the Legion environment

In this work we present the development of a parallel, multiblock flow solver using the Legion/Regent task-based programming framework. This work builds upon the Hypersonics Task-based Research (HTR) solver (Di Renzo et al., Comp. Phys. Comm. 2020), which is a highly scalable, compressible multi-species reacting Navier-Stokes solver designed for simulations of hypersonic turbulence and turbulent combustion. The novelty of this development lies in the treatment of general multi-block computational domains as a single logical instance (data layout in memory) composed by the union of grid blocks representing a complex computational domain. Our implementation leverages the capability of the Legion framework to manage complex index-spaces by automatically performing memory allocations and data synchronisation. This paradigm relieves the programmer and the user of the solver from manually setting up communication patterns among blocks and domain decomposition for multiprocessor computations. The correctness and scalability tests on both CPUs and GPUs are performed and demonstrated for canonical flow problems.