Problem
Research questions and friction points this paper is trying to address.
Develops AMReX for efficient multiphysics simulations.
Reduces computational cost with adaptive mesh refinement.
Supports diverse applications from laptops to exascale.
AMReX: Block-structured adaptive mesh refinement for multiphysics applications
🤖 AI Summary
Efficiently implementing block-structured adaptive mesh refinement (AMR) for multi-physics simulations—such as accelerator design, combustion, and cosmology—across heterogeneous hardware platforms (from laptops to exascale supercomputers) faces persistent challenges in numerical accuracy, scalability, and hardware adaptability. This paper introduces the first general-purpose AMR infrastructure unifying data containers, parallel iterators, and hardware-aware schedulers, enabling a unified software framework that supports seamless co-execution and heterogeneous deployment (CUDA, HIP, Kokkos, MPI, OpenMP) of PDE-based, particle-based, and particle-mesh coupled algorithms. Built upon a nested-grid abstraction, the framework decouples geometry, algorithmic logic, and backend execution. It achieves >90% strong and weak scaling efficiency on million-core systems, significantly reducing memory footprint and computational overhead. The infrastructure has already enabled over ten applications within the U.S. Exascale Computing Project (ECP).
📝 Abstract
Block-structured adaptive mesh refinement (AMR) provides the basis for the temporal and spatial discretization strategy for a number of Exascale Computing Project applications in the areas of accelerator design, additive manufacturing, astrophysics, combustion, cosmology, multiphase flow, and wind plant modeling. AMReX is a software framework that provides a unified infrastructure with the functionality needed for these and other AMR applications to be able to effectively and efficiently utilize machines from laptops to exascale architectures. AMR reduces the computational cost and memory footprint compared to a uniform mesh while preserving accurate descriptions of different physical processes in complex multiphysics algorithms. AMReX supports algorithms that solve systems of partial differential equations in simple or complex geometries and those that use particles and/or particle–mesh operations to represent component physical processes. In this article, we will discuss the core elements of the AMReX framework such as data containers and iterators as well as several specialized operations to meet the needs of the application projects. In addition, we will highlight the strategy that the AMReX team is pursuing to achieve highly performant code across a range of accelerator-based architectures for a variety of different applications.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Block-structured AMR for efficient multiphysics simulations
Unified framework for diverse applications and architectures
Optimized performance on exascale and accelerator-based systems
🔎 Similar Papers
No similar papers found.
