BoxLib with Tiling: An Adaptive Mesh Refinement Software Framework

Zhang, Weiqun; Almgren, Ann S.; Day, Marc; Nguyen, Tan; Shalf, John; Unat, Didem

doi:10.1137/15m102616x

Cited by 51 publications

(42 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This makes such a strategy suboptimal for use on many-core processors and GPUs. We have recently switched to a tiling approach where an OpenMP region is generated at the start of the hydrodynamics routine and the individual threads separately work on different partitions of each box (Zhang et al 2015). This results in much less overhead for the threading.…”

Section: Parallel Strategy and Performancementioning

confidence: 99%

White Dwarf Mergers on Adaptive Meshes. I. Methodology and Code Verification

Katz

Zingale

Calder

et al. 2016

ApJ

Self Cite

View full text Add to dashboard Cite

The Type Ia supernova (SN Ia) progenitor problem is one of the most perplexing and exciting problems in astrophysics, requiring detailed numerical modeling to complement observations of these explosions. One possible progenitor that has merited recent theoretical attention is the white dwarf (WD) merger scenario, which has the potential to naturally explain many of the observed characteristics of SNe Ia. To date there have been relatively few self-consistent simulations of merging WD systems using mesh-based hydrodynamics. This is the first paper in a series describing simulations of these systems using a hydrodynamics code with adaptive mesh refinement. In this paper we describe our numerical methodology and discuss our implementation in the compressible hydrodynamics code CASTRO, which solves the Euler equations, and the Poisson equation for self-gravity, and couples the gravitational and rotation forces to the hydrodynamics. Standard techniques for coupling gravitation and rotation forces to the hydrodynamics do not adequately conserve the total energy of the system for our problem, but recent advances in the literature allow progress and we discuss our implementation here. We present a set of test problems demonstrating the extent to which our software sufficiently models a system where large amounts of mass are advected on the computational domain over long timescales. Future papers in this series will describe our treatment of the initial conditions of these systems and will examine the early phases of the merger to determine its viability for triggering a thermonuclear detonation.

show abstract

Section: Parallel Strategy and Performancementioning

confidence: 99%

White Dwarf Mergers on Adaptive Meshes. I. Methodology and Code Verification

Katz

Zingale

Calder

et al. 2016

ApJ

Self Cite

View full text Add to dashboard Cite

show abstract

“…The first one is often referred to as patch-based or structured AMR, as originally proposed in Berger and Oliger (1984) and (Berger & Colella, 1989), and is the approach used in GeoClaw. It allows rectangular grid patches of arbitrary size and any integer refinement ratios between two level of grid patches (e.g., Adams et al, 2015;Bryan et al, 2014;Hornung & Kohn, 2002;Zhang et al, 2016). Another variant is cell-based AMR, which refines individual cells and often uses a quadtree or octree data structure to store the grid patch information.…”

Section: 1029/2019ms001635mentioning

confidence: 99%

Accelerating an Adaptive Mesh Refinement Code for Depth‐Averaged Flows Using GPUs

Qin

LeVeque

Motley

2019

J Adv Model Earth Syst

View full text Add to dashboard Cite

Solving the shallow water equations efficiently is critical to the study of natural hazards induced by tsunami and storm surge, since it provides more response time in an early warning system and allows more runs to be done for probabilistic assessment where thousands of runs may be required. Using adaptive mesh refinement speeds up the process by greatly reducing computational demands while accelerating the code using the graphics processing unit (GPU) does so through using faster hardware. Combining both, we present an efficient CUDA implementation of GeoClaw, an open source Godunov‐type high‐resolution finite volume numerical scheme on adaptive grids for shallow water system with varying topography. The use of adaptive mesh refinement and spherical coordinates allows modeling transoceanic tsunami simulation. Numerical experiments on the 2011 Japan tsunami and a local tsunami triggered by a hypothetical Mw 7.3 earthquake on the Seattle Fault illustrate the correctness and efficiency of the code, which implements a simplified dimensionally split version of the algorithms. Both numerical simulations are conducted on subregions on a sphere with adaptive grids that adequately resolve the propagating waves. The implementation is shown to be accurate and faster than the original when using Central Processing Units (CPUs) alone. The GPU implementation, when running on a single GPU, is observed to be 3.6 to 6.4 times faster than the original model running in parallel on a 16‐core CPU. Three metrics are proposed to evaluate relative performance of the model, which shows efficient usage of hardware resources.

show abstract

“…Like Kokkos, the metadata describing the layout of each array is carried throughout the program and into libraries, thereby offering a pathway to better library composability. TiDA is currently packaged as Fortran and C++ libraries and adopted by the BoxLib AMR framework [75].…”

Section: State Of the Artmentioning

confidence: 99%

Trends in Data Locality Abstractions for HPC Systems

Unat

Dubey

Hoefler

et al. 2017

IEEE Trans. Parallel Distrib. Syst.

Self Cite

View full text Add to dashboard Cite

Abstract-The cost of data movement has always been an important concern in high performance computing (HPC) systems. It has now become the dominant factor in terms of both energy consumption and performance. Support for expression of data locality has been explored in the past, but those efforts have had only modest success in being adopted in HPC applications for various reasons. However, with the increasing complexity of the memory hierarchy and higher parallelism in emerging HPC systems, locality management has acquired a new urgency. Developers can no longer limit themselves to low-level solutions and ignore the potential for productivity and performance portability obtained by using locality abstractions. Fortunately, the trend emerging in recent literature on the topic alleviates many of the concerns that got in the way of their adoption by application developers. Data locality abstractions are available in the forms of libraries, data structures, languages and runtime systems; a common theme is increasing productivity without sacrificing performance. This paper examines these trends and identifies commonalities that can combine various locality concepts to develop a comprehensive approach to expressing and managing data locality on future large-scale high-performance computing systems.

show abstract

BoxLib with Tiling: An Adaptive Mesh Refinement Software Framework

Cited by 51 publications

References 30 publications

White Dwarf Mergers on Adaptive Meshes. I. Methodology and Code Verification

White Dwarf Mergers on Adaptive Meshes. I. Methodology and Code Verification

Accelerating an Adaptive Mesh Refinement Code for Depth‐Averaged Flows Using GPUs

Trends in Data Locality Abstractions for HPC Systems

Contact Info

Product

Resources

About