11:10 AM - 11:55 AM (45 min)

AMReX

Donald Willcox (Lawrence Berkeley National Laboratory)

Abstract: AMReX is a software framework containing all the functionality to write massively parallel, block-structured adaptive mesh refinement (AMR) applications.

Key features of AMReX include:

• Primarily C++ with Fortran interfaces available
• Support for 1D, 2D and 3D
• Support for cell-centered, face-centered, edge-centered, and nodal data
• Support for particles and particle-mesh operations
• Support for embedded boundary (cut cell) representations of complex geometries
• Support for hyperbolic, parabolic, and elliptic solves on hierarchical adaptive grid structure
• Optional subcycling in time for time-dependent PDEs
• Hybrid parallelization strategy based on MPI+X, where X is OpenMP for multicore architectures, and primarily CUDA for architectures including NVIDIA GPUs and HIP for architectures including AMD GPUs
• Highly efficient parallel I/O, including native and hdf5 format
• Plotfile format supported by AmrVis, VisIt, ParaView, and yt.

Edit on GitHub