Hybrid Adaptive Ray-Moment Method (HARM2): A highly parallel method for radiation hydrodynamics on adaptive grids

Anna L. Rosen; Mark Krumholz; J. S Oishi; A. T Lee; Richard I. Klein

doi:10.1016/j.jcp.2016.10.048

Hybrid Adaptive Ray-Moment Method (HARM2): A highly parallel method for radiation hydrodynamics on adaptive grids

Anna L. Rosen, Mark Krumholz, J. S Oishi, A. T Lee, Richard I. Klein

Research output: Contribution to journal › Article › peer-review

Abstract

We present a highly-parallel multi-frequency hybrid radiation hydrodynamics algorithm that combines a spatially-adaptive long characteristics method for the radiation field from point sources with a moment method that handles the diffuse radiation field produced by a volume-filling fluid. Our Hybrid Adaptive Ray-Moment Method (HARM2) operates on patchbased adaptive grids, is compatible with asynchronous time stepping, and works with any moment method. In comparison to previous long characteristics methods, we have greatly improved the parallel performance of the adaptive long-characteristics method by developing a new completely asynchronous and non-blocking communication algorithm. As a result of this improvement, our implementation achieves near-perfect scaling up to O(103) processors on distributed memory machines. We present a series of tests to demonstrate the accuracy and performance of the method.

Original language	English
Pages (from-to)	924-942
Journal	Journal of Computational Physics
Volume	330
DOIs	https://doi.org/10.1016/j.jcp.2016.10.048
Publication status	Published - 2017

Access to Document

10.1016/j.jcp.2016.10.048

Cite this

@article{919627cd2f0e4b42b75ba41ae8e9ffec,

title = "Hybrid Adaptive Ray-Moment Method (HARM2): A highly parallel method for radiation hydrodynamics on adaptive grids",

abstract = "We present a highly-parallel multi-frequency hybrid radiation hydrodynamics algorithm that combines a spatially-adaptive long characteristics method for the radiation field from point sources with a moment method that handles the diffuse radiation field produced by a volume-filling fluid. Our Hybrid Adaptive Ray-Moment Method (HARM2) operates on patchbased adaptive grids, is compatible with asynchronous time stepping, and works with any moment method. In comparison to previous long characteristics methods, we have greatly improved the parallel performance of the adaptive long-characteristics method by developing a new completely asynchronous and non-blocking communication algorithm. As a result of this improvement, our implementation achieves near-perfect scaling up to O(103) processors on distributed memory machines. We present a series of tests to demonstrate the accuracy and performance of the method.",

author = "Rosen, {Anna L.} and Mark Krumholz and Oishi, {J. S} and Lee, {A. T} and Klein, {Richard I.}",

year = "2017",

doi = "10.1016/j.jcp.2016.10.048",

language = "English",

volume = "330",

pages = "924--942",

journal = "Journal of Computational Physics",

publisher = "Academic Press",

}

TY - JOUR

T1 - Hybrid Adaptive Ray-Moment Method (HARM2): A highly parallel method for radiation hydrodynamics on adaptive grids

AU - Rosen, Anna L.

AU - Krumholz, Mark

AU - Oishi, J. S

AU - Lee, A. T

AU - Klein, Richard I.

PY - 2017

Y1 - 2017

N2 - We present a highly-parallel multi-frequency hybrid radiation hydrodynamics algorithm that combines a spatially-adaptive long characteristics method for the radiation field from point sources with a moment method that handles the diffuse radiation field produced by a volume-filling fluid. Our Hybrid Adaptive Ray-Moment Method (HARM2) operates on patchbased adaptive grids, is compatible with asynchronous time stepping, and works with any moment method. In comparison to previous long characteristics methods, we have greatly improved the parallel performance of the adaptive long-characteristics method by developing a new completely asynchronous and non-blocking communication algorithm. As a result of this improvement, our implementation achieves near-perfect scaling up to O(103) processors on distributed memory machines. We present a series of tests to demonstrate the accuracy and performance of the method.

AB - We present a highly-parallel multi-frequency hybrid radiation hydrodynamics algorithm that combines a spatially-adaptive long characteristics method for the radiation field from point sources with a moment method that handles the diffuse radiation field produced by a volume-filling fluid. Our Hybrid Adaptive Ray-Moment Method (HARM2) operates on patchbased adaptive grids, is compatible with asynchronous time stepping, and works with any moment method. In comparison to previous long characteristics methods, we have greatly improved the parallel performance of the adaptive long-characteristics method by developing a new completely asynchronous and non-blocking communication algorithm. As a result of this improvement, our implementation achieves near-perfect scaling up to O(103) processors on distributed memory machines. We present a series of tests to demonstrate the accuracy and performance of the method.

U2 - 10.1016/j.jcp.2016.10.048

DO - 10.1016/j.jcp.2016.10.048

M3 - Article

VL - 330

SP - 924

EP - 942

JO - Journal of Computational Physics

JF - Journal of Computational Physics

ER -

Hybrid Adaptive Ray-Moment Method (HARM2): A highly parallel method for radiation hydrodynamics on adaptive grids

Abstract

Access to Document

Fingerprint

Cite this