An SN Algorithm for Modern Architectures

Baker, Randal Scott

doi:10.2172/1312619

Title: An S_N Algorithm for Modern Architectures

Full Record
Other Related Research

Abstract

LANL discrete ordinates transport packages are required to perform large, computationally intensive time-dependent calculations on massively parallel architectures, where even a single such calculation may need many months to complete. While KBA methods scale out well to very large numbers of compute nodes, we are limited by practical constraints on the number of such nodes we can actually apply to any given calculation. Instead, we describe a modified KBA algorithm that allows realization of the reductions in solution time offered by both the current, and future, architectural changes within a compute node.

Authors:

Baker, Randal Scott ^[1]

Los Alamos National Lab. (LANL), Los Alamos, NM (United States)

Publication Date:: 2016-08-29

Research Org.:: Los Alamos National Lab. (LANL), Los Alamos, NM (United States)

Sponsoring Org.:: USDOE National Nuclear Security Administration (NNSA)

OSTI Identifier:: 1312619

Report Number(s):: LA-UR-16-26593

Grant/Contract Number:: AC52-06NA25396

Resource Type:: Accepted Manuscript

Country of Publication:: United States

Language:: English

Subject:: 97 MATHEMATICS AND COMPUTING

Citation Formats


                    Baker, Randal Scott. An SN Algorithm for Modern Architectures.  United States: N. p., 2016. 
Web.  doi:10.2172/1312619.

Copy to clipboard


                    Baker, Randal Scott. An SN Algorithm for Modern Architectures.  United States.  https://doi.org/10.2172/1312619

Copy to clipboard


                    Baker, Randal Scott. Mon .  
"An SN Algorithm for Modern Architectures".  United States.  https://doi.org/10.2172/1312619.  https://www.osti.gov/servlets/purl/1312619.

Copy to clipboard


                    
@article{osti_1312619,

  title        = {An SN Algorithm for Modern Architectures},

  author       = {Baker, Randal Scott},

  abstractNote = {LANL discrete ordinates transport packages are required to perform large, computationally intensive time-dependent calculations on massively parallel architectures, where even a single such calculation may need many months to complete. While KBA methods scale out well to very large numbers of compute nodes, we are limited by practical constraints on the number of such nodes we can actually apply to any given calculation. Instead, we describe a modified KBA algorithm that allows realization of the reductions in solution time offered by both the current, and future, architectural changes within a compute node.},

  doi          = {10.2172/1312619},

  journal      = {},

  number       = ,

  volume       = ,

  place        = {United States},

  year         = {2016},

  month        = {8}

}

Copy to clipboard

Journal Article:

Free Publicly Available Full Text

Accepted Manuscript (DOE)

Publisher's Version of Record

https://doi.org/10.2172/1312619

Other availability

Search WorldCat to find libraries that may hold this journal

Save / Share:

Export Metadata

Save to My Library

Similar Records in DOE PAGES and OSTI.GOV collections:

A Novel Algorithm for Solving the Multidimensional Neutron Transport Equation on Massively Parallel Architectures

Technical Report Azmy, Yousry

We employ the Integral Transport Matrix Method (ITMM) as the kernel of new parallel solution methods for the discrete ordinates approximation of the within-group neutron transport equation. The ITMM abandons the repetitive mesh sweeps of the traditional source iterations (SI) scheme in favor of constructing stored operators that account for the direct coupling factors among all the cells' fluxes and between the cells' and boundary surfaces' fluxes. The main goals of this work are to develop the algorithms that construct these operators and employ them in the solution process, determine the most suitable way to parallelize the entire procedure, andmore »« less
https://doi.org/10.2172/1154732

Full Text Available
Parallel S/sub n/ transport algorithms

Journal Article Wienke, B R ; Hiromoto, R E - Transp. Theory Stat. Phys.; (United States)

Parallelization of standard multigroup methods used to solve the linear (Boltzmann) transport equation in the discrete ordinates (S/sub n/) representation and coupled chaotic iteraction schemes are the focus of this analysis. On the Denelcor HEP, we extend two serial iteration schemes, categorize speedup, and contrast ordered and chaotic methods. Ordered and chaotic iteration strategies, with and without acceleration, support relatively unrestricted parallelism and appear to be robust parallel techniques. Parallel modifications and recording efforts to serial iteration algorithms are straightforward, actual speedup and efficiency are high, and payoff appears substantial, largely due to the coarse computational granularity of the multigroupmore »« less
https://doi.org/10.1080/00411458608210444
Parallel S{sub n} Sweeps on Unstructured Grids: Algorithms for Prioritization, Grid Partitioning, and Cycle Detection

Journal Article Plimpton, Steven J ; Hendrickson, Bruce ; Burns, Shawn P ; ... - Nuclear Science and Engineering

The method of discrete ordinates is commonly used to solve the Boltzmann transport equation. The solution in each ordinate direction is most efficiently computed by sweeping the radiation flux across the computational grid. For unstructured grids this poses many challenges, particularly when implemented on distributed-memory parallel machines where the grid geometry is spread across processors. We present several algorithms relevant to this approach: (a) an asynchronous message-passing algorithm that performs sweeps simultaneously in multiple ordinate directions, (b) a simple geometric heuristic to prioritize the computational tasks that a processor works on, (c) a partitioning algorithm that creates columnar-style decompositions formore »« less
https://doi.org/10.13182/NSE150-267
Iterative methods for solving x-y geometry S[sub N] problems on parallel architecture computers

Journal Article Yavuz, M ; Larsen, E W - Nuclear Science and Engineering; (United States)

In this paper, geometric domain decomposition methods are described for solving x-y geometry discrete ordinates (S[sub N]) problems on parallel architecture computers. First, a parallel source iteration scheme is developed; here, one subdivides the spatial domain of the problem, performs transport sweeps independently in each subdomain, and iterates on the scattering source and the interface fluxes between each subdomain. Second, a parallel diffusion synthetic acceleration (DSA) scheme is developed to speed up the convergence of the parallel source iteration. These schemes have been implemented on the IBM RP3, a shared/distributed memory parallel computer. The numerical results show that the parallelmore »« less
Software architecture of the light weight kernel, catamount.

Conference Kelly, Suzanne Marie

Catamount is designed to be a low overhead operating system for a parallel computing environment. Functionality is limited to the minimum set needed to run a scientific computation. The design choices and implementations will be presented. A massively parallel processor (MPP), high performance computing (HPC) system is particularly sensitive to operating system overhead. Traditional, multi-purpose, operating systems are designed to support a wide range of usage models and requirements. To support the range of needs, a large number of system processes are provided and are often interdependent on each other. The overhead of these processes leads to an unpredictable amountmore »« less

Similar Records

Title: An SN Algorithm for Modern Architectures

Abstract

Citation Formats

Title: An S_N Algorithm for Modern Architectures