Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Parallel algorithms.”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 487 records · Page 27

Computational Fluid Dynamics Using the Adaptive Wavelet-Collocation Method

Advancements to the adaptive wavelet-collocation method over the last decade have opened up a number of new possible areas for active research. Volume penalization techniques allow complex immersed boundary conditions to be used with high efficiency for both internal and external flows. Anisotropic methods make it possible to use body-fitted meshes while still taking advantage of the dynamic adaptability properties wavelet-based methods provide. The parallelization of the approach has made it possible to perform large high-resolution simulations of detonation initiation and fluid instabilities to uncover new physical insights that would otherwise be difficult to discover. Other developments include space-time adaptive methods and nonreflecting boundary conditions. This article summarizes the work performed using the adaptive wavelet-collocation method developed by Vasilyev and coworkers over the past decade.

42 ENGINEERING↗

PyFLEXTRKR: a flexible feature tracking Python software for convective cloud analysis

Abstract. This paper describes the new open-source framework PyFLEXTRKR (Python FLEXible object TRacKeR), a flexible atmospheric feature tracking software package with specific capabilities to track convective clouds from a variety of observations and model simulations. This software can track any atmospheric 2D objects and handle merging and splitting explicitly. The package has a collection of multi-object identification algorithms, scalable parallelization options, and has been optimized for large datasets including global high-resolution data. We demonstrate applications of PyFLEXTRKR on tracking individual deep convective cells and mesoscale convective systems from observations and model simulations ranging from large-eddy resolving (∼100s m) to mesoscale (∼10s km) resolutions. Visualization, post-processing, and statistical analysis tools are included in the package. New Lagrangian analyses of convective clouds produced by PyFLEXTRKR applicable to a wide range of datasets and scales facilitate advanced model evaluation and development efforts as well as scientific discovery.

54 ENVIRONMENTAL SCIENCES↗

QUEST: QUantum Electron Simulation Toolbox

QUEST is a part of the SciDAC project on next generation multi-scale quantum simulation software for strongly correlated materials. It is a Fortran 90/95 package that implements the determinant quantum Monte Carlo (DQMC) method for simulation of magnetic, superconducting, and metal-insulator transitions in model Hamiltonians. In this paper, we show how QUEST is capable of treating lattices of unprecedentedly large sizes and how this can be fruitful in the study of the physics of trapped fermionic system, in the development of more efficient solvers for Dynamical Mean Field Theory (DMFT) and as a tool to test and, in the future, improve diagrammatic approaches such as the Parquet approximation. We will also present a range of synergistic activities on the development of stable and robust numerical algorithms and hybrid granularity parallelization scheme that combines algorithmic and implementation techniques to high-performance DQMC simulation. The work reported here is a key step forward in achieving the goals of our SciDAC project.

97 MATHEMATICS AND COMPUTING↗

Research on Fast-Doppler-Broadening of neutron cross sections

A Fast-Doppler-Broadening method is developed in this work to broaden Continuous Energy neutron cross-sections for Monte Carlo calculations. Gauss integration algorithm and parallel computing are implemented in this method, which is unprecedented in the history of cross section processing. Compared to the traditional code (NJOY, SIGMA1, etc.), the new Fast-Doppler-Broadening method shows a remarkable speedup with keeping accuracy. The purpose of using Gauss integration is to avoid complex derivation of traditional broadening formula and heavy load of computing complementary error function that slows down the Doppler broadening process. The OpenMP environment is utilized in parallel computing which can take full advantage of modern multi-processor computers. Combination of the two can reduce processing time of main actinides (such as {sup 238}U, {sup 235}U) to an order of magnitude of 1{approx}2 seconds. This new method is fast enough to be applied to Online Doppler broadening. It can be combined or coupled with Monte Carlo transport code to solve temperature dependent problems and neutronics-thermal hydraulics coupled scheme which is a big challenge for the conventional NJOY-MCNP system. Examples are shown to determine the efficiency and relative errors compared with the NJOY results. A Godiva Benchmark is also used in order to test the ACE libraries produced by the new method. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Development of a High Fidelity System Analysis Code for Generation IV Reactors

Traditional nuclear reactor system analysis codes such as RELAP and TRAC employ an operator split methodology. In this approach, each of the physics (fluid flow, heat conduction and neutron diffusion) is solved separately and the coupling terms are done explicitly. This approach limits accuracy (first order in time at best) and makes the codes slow in running since the explicit coupling imposes stability restrictions on the time step size. These codes have been extensively tested and validated for the existing LWRs. However, for GEN IV nuclear reactor designs which tend to have long lasting transients resulting from passive safety systems, the performance is questionable and modern high fidelity simulation tools will be required. The requirement for accurate predictability is the motivation for a large scale overhaul of all of the models and assumptions in transient nuclear reactor safety simulation software. At INL we have launched an effort with the long term goal of developing a high fidelity system analysis code that employs modern physical models, numerical methods, and computer science for transient safety analysis of GEN IV nuclear reactors. Modern parallel solution algorithms will be employed through utilizing the nonlinear solution software package PETSc developed by Argonne National Laboratory. The physical models to be developed will have physically realistic length scales and time scales. The solution algorithm will be based on the physics-based preconditioned Jacobian-free Newton-Krylov solution methods. In this approach all of the physical models are solved implicitly and simultaneously in a single nonlinear system. This includes the coolant flow, nonlinear heat conduction, neutron kinetics, and thermal radiation, etc. Including modern physical models and accurate space and time discretizations will allow the simulation capability to be second order accurate in space and in time. This paper presents the current status of the development efforts as well as some results from analyzing a simplified primary system model of GNEP’s advanced burner test reactor (ABTR) designed by Argonne. Various transient analyses are performed with this simplified ABTR model to study two fundamental issues related to system analysis codes – accuracy of numeric algorithm and efficiency. The accuracy study is carried by comparing the second order method with the first order method. The results show that numerical errors in the first order method are large and it is very difficult to distinguish numerical errors from physical modeling errors. On the other hand, second order method yields small numerical errors and it is very easy to spot physical modeling errors. The efficiency study is carried out by comparing the time steps for the fully implicit solution algorithm versus CFL stability limit methods. The dynamic time steps used in a fully implicit method will adjust the time step to resolve the time scale during the various stages of a long lasting transient. This will make a computer code based on fully implicit methods run more efficiently versus a CFL stability limit method code like RELAP, in which a particle of fluid cannot cross a control volume in a single time step.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

OpenMP Parallelization and Optimization of Graph-Based Machine Learning Algorithms

In this paper, we investigate the OpenMP parallelization and optimization of two novel data classification algorithms. The new algorithms are based on graph and PDE solution techniques and provide significant accuracy and performance advantages over traditional data classification algorithms in serial mode. The methods leverage the Nystrom extension to calculate eigenvalue/eigenvectors of the graph Laplacian and this is a self-contained module that can be used in conjunction with other graph-Laplacian based methods such as spectral clustering. We use performance tools to collect the hotspots and memory access of the serial codes and use OpenMP as the parallelization language to parallelize the most time-consuming parts. Where possible, we also use library routines. We then optimize the OpenMP implementations and detail the performance on traditional supercomputer nodes (in our case a Cray XC30), and test the optimization steps on emerging testbed systems based on Intel’s Knights Corner and Landing processors. We show both performance improvement and strong scaling behavior. Finally, a large number of optimization techniques and analyses are necessary before the algorithm reaches almost ideal scaling.

97 MATHEMATICS AND COMPUTING↗

Mechanisms for the convergence of time-parallelized, parareal turbulent plasma simulations

Parareal is a recent algorithm able to parallelize the time dimension in spite of its sequential nature. It has been applied to several linear and nonlinear problems and, very recently, to a simulation of fully-developed, two-dimensional drift wave turbulence. The mere fact that parareal works in such a turbulent regime is in itself somewhat unexpected, due to the characteristic sensitivity of turbulence to any change in initial conditions. This fundamental property of any turbulent system should render the iterative correction procedure characteristic of the parareal method inoperative, but this seems not to be the case. In addition, the choices that must be made to implement parareal (division of the temporal domain, election of the coarse solver and so on) are currently made using trial-and-error approaches. Here, we identify the mechanisms responsible for the convergence of parareal of these simulations of drift wave turbulence. We also investigate which conditions these mechanisms impose on any successful parareal implementation. The results reported here should be useful to guide future implementations of parareal within the much wider context of fully-developed fluid and plasma turbulent simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Distributed Augmentation, Hypersweeps, and Branch Decomposition of Contour Trees for Scientific Exploration

Contour trees describe the topology of level sets in scalar fields and are widely used in topological data analysis and visualization. A main challenge of utilizing contour trees for large-scale scientific data is their computation at scale using highperformance computing. To address this challenge, recent work has introduced distributed hierarchical contour trees for distributed computation and storage of contour trees. However, effective use of these distributed structures in analysis and visualization requires subsequent computation of geometric properties and branch decomposition to support contour extraction and exploration. In this work, we introduce distributed algorithms for augmentation, hypersweeps, and branch decomposition that enable parallel computation of geometric properties, and support the use of distributed contour trees as query structures for scientific exploration. Finally, we evaluate the parallel performance of these algorithms and apply them to identify and extract important contours for scientific visualization.

97 MATHEMATICS AND COMPUTING↗

Parallel peak pruning for scalable SMP contour tree computation

As data sets grow to exascale, automated data analysis and visualisation are increasingly important, to intermediate human understanding and to reduce demands on disk storage via in situ analysis. Trends in architecture of high performance computing systems necessitate analysis algorithms to make effective use of combinations of massively multicore and distributed systems. One of the principal analytic tools is the contour tree, which analyses relationships between contours to identify features of more than local importance. Unfortunately, the predominant algorithms for computing the contour tree are explicitly serial, and founded on serial metaphors, which has limited the scalability of this form of analysis. While there is some work on distributed contour tree computation, and separately on hybrid GPU-CPU computation, there is no efficient algorithm with strong formal guarantees on performance allied with fast practical performance. Here in this paper, we report the first shared SMP algorithm for fully parallel contour tree computation, withfor-mal guarantees of O(lgnlgt) parallel steps and O(n lgn) work, and implementations with up to 10x parallel speed up in OpenMP and up to 50x speed up in NVIDIA Thrust.

97 MATHEMATICS AND COMPUTING↗

A Distributed-Memory Package for Dense Hierarchically Semi-Separable Matrix Computations Using Randomization

In this paper, we present a distributed-memory library for computations with dense structured matrices. A matrix is considered structured if its off-diagonal blocks can be approximated by a rank-deficient matrix with low numerical rank. Here, we use Hierarchically Semi-Separable (HSS) representations. Such matrices appear in many applications, for example, finite-element methods, boundary element methods, and so on. Exploiting this structure allows for fast solution of linear systems and/or fast computation of matrix-vector products, which are the two main building blocks of matrix computations. The compression algorithm that we use, that computes the HSS form of an input dense matrix, relies on randomized sampling with a novel adaptive sampling mechanism. We discuss the parallelization of this algorithm and also present the parallelization of structured matrix-vector product, structured factorization, and solution routines. The efficiency of the approach is demonstrated on large problems from different academic and industrial applications, on up to 8,000 cores. Finally, this work is part of a more global effort, the STRUctured Matrices PACKage (STRUMPACK) software package for computations with sparse and dense structured matrices. Hence, although useful on their own right, the routines also represent a step in the direction of a distributed-memory sparse solver.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

Phlex: Parallel, Hierarchical, and Layered EXecution of data-processing algorithms

Phlex is a computing framework supporting the parallel, hierarchical, and layered execution of data-processing algorithms. It is based on the functional-programming paradigm, thus guaranteeing thread-safety when invoking user-defined pure functions. Phlex allows users to specify arbitrary graph-based hierarchies of data organization, enabling more flexible processing of data as required by the constraints of the program.

Knoepfel, KyleJ. [Fermi National Accelerator Labor↗

A PETSc Parallel-in-Time Solver based on MGRIT algorithm

We address the development of a modular implementation of the MGRIT (MultiGrid-In-Time) algorithm to solve linear and nonlinear systems that arise from the discretization of evolutionary models with a parallel-in-time approach in the context of the PETSc (the Portable, Extensible Toolkit for Scientific computing) library. Our aim is to give the opportunity of predicting the performance gain achievable when using the MGRIT approach instead of the Time Stepping integrator (TS). To this end, we analyze the performance parameters of the algorithm that provide a-priori the best number of processing elements and grid levels to use to address the scaling of MGRIT, regarded as a parallel iterative algorithm proceeding along the time dimension.

Mele, Valeria↗

A Cut-Cell Approach for 2D Cartesian Meshes that Preserves Orthogonal Grid Sweep Ordering

In this paper, we present a cut-cell methodology for solving the two-dimensional neutral-particle transport equation on an orthogonal Cartesian grid. We allow the rectangular cell to be subdivided into two polygonal subcells. We ensure that this division (or cut) conserves the volumes of the materials in the subcells and we utilize a step-characteristics (SC) slice balance approach (SBA) to calculate the angular fluxes exiting the cell as well as the average scalar fluxes in each subcell. Solving the discrete ordinates transport equation on an arbitrary mesh has historically been difficult to parallelize while maintaining good parallel efficiency. However on Cartesian meshes, the KBA algorithm maintains good parallel efficiency using a direct solve. The ability to preserve this algorithm was a driving factor in the development of our cut-cell method. This method also provides a more accurate depiction of a material interface in a cell, which leads to more accurate solutions downstream of this cell. As a result, fewer spatial cells can be utilized, resulting in reduced memory requirements. We apply this approach in the 2D/3D discrete ordinates neutral-particle transport code Denovo, where we analyze a 2D 3 x 3 lattice of pincells. We show that, for eigenvalue problems, a significant increase in accuracy for a given mesh size is gained by utilizing the cut-cell, SC equations instead of the standard homogenized-cell, SC equations.

99 GENERAL AND MISCELLANEOUS↗

Algorithmic Improvements for Portable Event-Based Monte Carlo Transport Using the Nvidia Thrust Library

High performance computing environments are progressively moving towards many-core computing architectures. The Los Alamos National Laboratory Trinity machine, available in late 2016, will use both Intel Xeon Haswell processors and Intel Xeon Phi Knights Landing many integrated core (MIC) coprocessors. The Lawrence Livermore National Laboratory Sierra machine, available in 2018, will use an IBM PowerPC architecture along with Nvidia graphics processing units (GPUs). Applications that must work in this supercomputing environment must continue to adapt in order to take advantage of the diverse hardware architectures that are coming. A significant consideration is not only the performance of the application on a given platform but also the portability of the application to other platforms. The algorithmic improvements presented in this paper build upon recently-reported work on event-based Monte Carlo transport in the ALPSMC code that models particle transport in one-dimensional binary stochastic media. That paper discussed the lack of available vectorization in the traditional history-based algorithm used for Monte Carlo transport and presented a data parallel event-based algorithm implemented using the Nvidia Thrust library for portability. The performance of the data parallel event-based algorithm implemented using Thrust was compared to a native CUDA implementation. The conclusions from that work were that the Thrust library abstraction technique caused too significant a loss in performance but that the event-based method was a viable option that should be further investigated. In this paper, we describe algorithmic improvements to the data parallel event-based algorithm previously presented. We made further algorithmic optimizations to the event-based CUDA implementation, most notably: data structure changes, a new conditional particle removal scheme in the event-based process, and the use of multiple GPUs. In addition to improvements to the algorithm, we re-implemented the Thrust version from the now further optimized CUDA version, giving a greater chance for success at a performant abstraction. Finally, we revisited our previous assumptions about the inability of the history-based method to achieve performance on vector style architectures such as the MICs and GPUs, with surprising and promising results. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Structural properties of small rhodium clusters

We report a systematic study of the structural properties of rhodium clusters at the atomistic level. A novel global-minimum search algorithm, known as parallel tempering multicanonical basin hopping plus genetic algorithm (PTMBHGA), is used to obtain the geometrical structures with lowest minima at the semi-empirical level where Gupta potential is used to describe the atomic interaction among the rhodium atoms. These structures are then re-optimized at the density functional theory (DFT) level with exchange-correlation energy approximated by Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation (GGA). The structures are optimized for different spin multiplicities. The ones with lowest energies will be taken as ground-state structures. In most cases, we observe only minor changes in the geometry and bond length of the clusters as a result of DFT-level re-optimization. Only in some limited cases, the initial geometries obtained from the PTMBHGA are modified by the re-optimization. The variation of structural properties, such as ground-state geometry, symmetry and binding energy, with respect to the cluster size is studied and agreed well with other results available in the literature.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗