Skip to main content Skip to secondary navigation
Construction alert

Under construction

This website is under construction. For lab group information, please contact lele@stanford.edu 

AMR-H (work in progress)

Main content start
 

This work is supported by the U.S. National Science Foundation (NSF-OAC-2103509)

AMR-H is a high performance computational infrastructure for compressible flows simulations. The numerical schemes are based on high-order finite difference methods on a structured adaptive mesh refinement (AMR) framework. AMR-H is designed for efficient large scale computations on multiple high performance computing (HPC) platforms. The framework is developed following data-centric parallelism, which is achieved by Legion programing model. The base-level code is particularly optimized for the state-of-the-art HPC architectures. Additionally, for a general computational platform, the performance portability is further gained by Kokkos programing model. AMR-H is primarily implemented in C++, which is allowed to be further extended for various applications for multi-physics computations.

 

The code development team is looking for collaborations at any stage. If interested please contact the project PI.