ForestClaw: Hybrid forest-of-octrees AMR for hyperbolic conservation laws

  title={ForestClaw: Hybrid forest-of-octrees AMR for hyperbolic conservation laws},
  author={Carsten Burstedde and Donna A. Calhoun and Kyle T. Mandli and Andy R. Terrel},
We present a new hybrid paradigm for parallel adaptive mesh refinement (AMR) that combines the scalability and lightweight architecture of tree-based AMR with the computational efficiency of patch-based solvers for hyperbolic conservation laws. The key idea is to interpret each leaf of the AMR hierarchy as one uniform compute patch in $\sR^d$ with $m^d$ degrees of freedom, where $m$ is customarily between 8 and 32. Thus, computation on each patch can be optimized for speed, while we inherit the… 

Figures and Tables from this paper

p 4 est : Scalable Algorithms for Parallel Adaptive Mesh Refinement
The scalability of the p4est code for parallel adaptive mesh refinement (AMR) is examined, which implements several algorithms to create a dynamic distributed mesh data structure, to refine, coarsen, and 2:1 balance it, and to repartition it between the parallel processes.
Parallel Memory-Efficient Adaptive Mesh Refinement on Structured Triangular Meshes with Billions of Grid Cells
We present sam(oa)2, a software package for a dynamically adaptive, parallel solution of 2D partial differential equations on triangular grids created via newest vertex bisection. An element order
Coarse Mesh Partitioning for Tree-Based AMR
An algorithm to partition the elements of the coarse mesh such that the fine mesh can be load-balanced to equal element counts per process regardless of the element-to-tree map, and each process that holds fine mesh elements has access to the meta data of all relevant trees.
A Generic Interface for Godunov-Type Finite Volume Methods on Adaptive Triangular Meshes
This interface requires application developers only to provide problem-specific implementations of a set of operators, while sam(oa)\(^2\) transparently manages HPC features such as memory-efficient adaptive mesh refinement, parallelism in distributed and shared memory and vectorization of Riemann solvers.
Extreme-Scale Block-Structured Adaptive Mesh Refinement
This article proposes to exploit the hierarchical nature of the block-structured domain partitioning by creating a lightweight, temporary copy of the core data structure that acts as a local and fully distributed proxy data structure for the dynamic load balancing.
Load Balancing and Patch-Based Parallel Adaptive Mesh Refinement for Tsunami Simulation on Heterogeneous Platforms Using Xeon Phi Coprocessors
A patch-based approach for tsunami simulation with parallel adaptive mesh refinement on the Salomon supercomputer, using vectorised versions of the approximate Riemann solvers, and finding that relatively small patches with 82 cells resulted in the smallest execution times.
Massively Parallel Algorithms for the Lattice Boltzmann Method on NonUniform Grids
This article presents parallel algorithms, distributed data structures, and communication routines that are implemented in the software framework waLBerla in order to support large-scale, massively parallel lattice Boltzmann-based simulations on nonuniform grids, and evaluates the performance on two current petascale supercomputers.
Accelerating wave-propagation algorithms with adaptive mesh refinement using the Graphics Processing Unit (GPU)
Recently, capabilities to accelerate the code by using the Graphics Process Unit (GPU) are added and Routines that manage CPU and GPU AMR data and facilitate the execution of GPU kernels are added.


Parallel geometric-algebraic multigrid on unstructured forests of octrees
A parallel multigrid method for solving variable-coefficient elliptic partial differential equations on arbitrary geometries using highly adapted meshes that are built from an unstructured hexa-hedral macro mesh, in which each macro element is adaptively refined as an octree.
p4est: Scalable Algorithms for Parallel Adaptive Mesh Refinement on Forests of Octrees
We present scalable algorithms for parallel adaptive mesh refinement and coarsening (AMR), partitioning, and 2:1 balancing on computational domains composed of multiple connected two-dimensional
ALPS: A framework for parallel adaptive PDE solution
ALPS, a library for dynamic mesh adaptation of PDEs that is designed to scale to hundreds of thousands of compute cores, is developed, which supports arbitrary-order accurate continuous and discontinuous finite element/spectral element discretizations on general geometries.
Nodal discontinuous Galerkin methods on graphics processors
Low-Cost Parallel Algorithms for 2:1 Octree Balance
A lightweight collective algorithm for reversing the asymmetric communication pattern induced by non-local octant interactions is proposed and implemented as part of the opensource “p4est” software.
Extreme-Scale AMR
New parallel algorithms for parallel dynamic AMR on forest-ofoctrees geometries with arbitrary-order continuous and discontinuous finite/spectral element discretizations for multiscale geophysics problems are presented.
Towards adaptive mesh PDE simulations on petascale computers
Although supercomputing performance has outpaced Moore’s Law over the past several decades due to increased concurrency, the curse of dimensionality imposes much slower scientific returns; e.g. improvement in mesh resolution grows at best with the onefourth power of peak performance in the case of evolutionary PDEs.
Terascale Spectral Element Algorithms and Implementations
The development and implementation of an efficient spectral element code for multimillion gridpoint simulations of incompressible flows in general two- and three-dimensional domains is described and a stabilization scheme that admits full use of the method’s high-order accuracy is described.
Logically Rectangular Grids and Finite Volume Methods for PDEs in Circular and Spherical Domains
We describe a class of logically rectangular quadrilateral and hexahedral grids for solving PDEs in circular and spherical domains, including grid mappings for the circle, the surface of the sphere,
ManyClaw: Slicing and dicing Riemann solvers for next generation highly parallel architectures
The goal is to separate the low level parallelism and the physical equations thus providing users the capability to leverage intra-node parallelism without explicitly writing code to take advantage of newer architectures.