Adaptable Poly-Engineering Simulator (APES)

APES is a suite of mesh based simulation tools built around an octree mesh infrastructure. It provides various steps required for the simulation on large scale parallel computing facilities from the mesh generation to post-processing.

The applications are MPI parallel and scale well to many thousands of processes. There are currently two solvers developed in the suite and available under permissive open-source licenses:

  • Ateles is a high-order Discontinuous Galerkin solver that employs Legendre polynomials in cubical elements as basis functions.
  • Musubi is a multi-level Lattice-Boltzmann solver that implements efficient kernels and the possibility to simulate multiple species with the Maxwell-Stefan model.

The mesh generator Seeder produces octree meshes and reads STL data to describe geometries. The resulting mesh data is simple and designed for efficient parallel reading on distributed memory systems. The core library implementing the handling of these octree meshes is implemented in TreElM.

The solvers write their data in a native format on disk with MPI-IO, and come with dedicated tools to post-process this data further into VTK files. All tools are using Lua scripts as user interfaces, which enables a great flexibility in the definition of simulation setups. This functionality is provided by Aotus.

Apes, or more generally simians are:

  • adapting to given tasks
  • living in trees
  • relatively independent
  • working in teams
  • smart
STS