Portably performant discontinuous Galerkin adaptive mesh library
DGTile is a lightweight C++17 adaptive mesh library is meant to support explicit discontinuous Galerkin applications on high performance computing machines. The mesh supports block-structured adaptivity (isotropic refinement and coarsening), where each block represents a Cartesian grid in one, two and three spatial dimensions. Additionally, support for modal discontinuous Galerkin basis functions is provided. Distributed memory parallelism (MPI) is achieved by partitioning blocks over MPI ranks, and shared memory parallelism is achieved by parallelizing functors over each block.
DGTile has required dependencies on:
DGTile has an optional dependency on:
if -DBUILD_TESTING=ON.
At Sandia, DGTile is SCR 2806.0