Publications

Abstract not provided.

We present an algorithm to maintain the connected components of a graph that arrives as an infinite stream of edges. We formalize the algorithm on X-stream, a new parallel theoretical computational model for infinite streams. Connectivity-related queries, including component spanning trees, are supported with some latency, returning the state of the graph at the time of the query. Because an infinite stream may eventually exceed the storage limits of any number of finite-memory processors, we assume an aging command or daemon where "uninteresting" edges are removed when the system nears capacity. Following an aging command the system will block queries until its data structures are repaired, but edges will continue to be accepted from the stream, never dropped. The algorithm will not fail unless a model-specific constant fraction of the aggregate memory across all processors is full. In normal operation, it will not fail unless aggregate memory is completely full. Unlike previous theoretical streaming models designed for finite graphs that assume a single shared memory machine or require arbitrary-size intemediate files, X-stream distributes a graph over a ring network of finite-memory processors. Though the model is synchronous and reminiscent of systolic algorithms, our implementation uses an asynchronous message-passing system. We argue the correctness of our X-stream connected components algorithm, and give preliminary experimental results on synthetic and real graph streams.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

We discuss the relative complexity and computational cost of several popular many-body empirical potentials, developed by the materials science community over the past 30 years. The inclusion of more detailed many-body effects has come at a computational cost, but the cost still scales linearly with the number of atoms modeled. This is enabling very large molecular dynamics simulations with unprecedented atomic-scale fidelity to physical and chemical phenomena. The cost and scalability of the potentials, run in serial and parallel, are benchmarked in the LAMMPS molecular dynamics code. Several recent large calculations performed with these potentials are highlighted to illustrate what is now possible on current supercomputers. We conclude with a brief mention of high-performance computing architecture trends and the research issues they raise for continued potential development and use. © 2012 Materials Research Society.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Peridynamics is a nonlocal extension of classical continuum mechanics. The discrete peridynamic model has the same computational structure as a molecular dynamics model. This document provides a brief overview of the peridynamic model of a continuum, then discusses how the peridynamic model is discretized within LAMMPS. An example problem is also included.

We describe a parallel library written with message-passing (MPI) calls that allows algorithms to be expressed in the MapReduce paradigm. This means the calling program does not need to include explicit parallel code, but instead provides "map" and "reduce" functions that operate independently on elements of a data set distributed across processors. The library performs needed data movement between processors. We describe how typical MapReduce functionality can be implemented in an MPI context, and also in an out-of-core manner for data sets that do not fit within the aggregate memory of a parallel machine. Our motivation for creating this library was to enable graph algorithms to be written as MapReduce operations, allowing processing of terabyte-scale data sets on traditional MPI-based clusters. We outline MapReduce versions of several such algorithms: vertex ranking via PageRank, triangle finding, connected component identification, Luby's algorithm for maximally independent sets, and single-source shortest-path calculation. To test the algorithms on arbitrarily large artificial graphs we generate randomized R-MAT matrices in parallel; a MapReduce version of this operation is also described. Performance and scalability results for the various algorithms are presented for varying size graphs on a distributed-memory cluster. For some cases, we compare the results with non-MapReduce algorithms, different machines, and different MapReduce software, namely Hadoop. Our open-source library is written in C++, is callable from C++, C, Fortran, or scripting languages such as Python, and can run on any parallel platform that supports MPI. © 2011 Elsevier B.V. All rights reserved.

Abstract not provided.

Many of the most important and hardest-to-solve problems related to the synthesis, performance, and aging of materials involve diffusion through the material or along surfaces and interfaces. These diffusion processes are driven by motions at the atomic scale, but traditional atomistic simulation methods such as molecular dynamics are limited to very short timescales on the order of the atomic vibration period (less than a picosecond), while macroscale diffusion takes place over timescales many orders of magnitude larger. We have completed an LDRD project with the goal of developing and implementing new simulation tools to overcome this timescale problem. In particular, we have focused on two main classes of methods: accelerated molecular dynamics methods that seek to extend the timescale attainable in atomistic simulations, and so-called 'equation-free' methods that combine a fine scale atomistic description of a system with a slower, coarse scale description in order to project the system forward over long times.

Abstract not provided.

Abstract not provided.

Evaporation and condensation at a liquid/vapor interface are ubiquitous interphase mass and energy transfer phenomena that are still not well understood. We have carried out large scale molecular dynamics simulations of Lennard-Jones (LJ) fluids composed of monomers, dimers, or trimers to investigate these processes with molecular detail. For LJ monomers in contact with a vacuum, the evaporation rate is found to be very high with significant evaporative cooling and an accompanying density gradient in the liquid domain near the liquid/vapor interface. Increasing the chain length to just dimers significantly reduces the evaporation rate. We confirm that mechanical equilibrium plays a key role in determining the evaporation rate and the density and temperature profiles across the liquid/vapor interface. The velocity distributions of evaporated molecules and the evaporation and condensation coefficients are measured and compared to the predictions of an existing model based on kinetic theory of gases. Our results indicate that for both monatomic and polyatomic molecules, the evaporation and condensation coefficients are equal when systems are not far from equilibrium and smaller than one, and decrease with increasing temperature. For the same reduced temperature TT c, where Tc is the critical temperature, these two coefficients are higher for LJ dimers and trimers than for monomers, in contrast to the traditional viewpoint that they are close to unity for monatomic molecules and decrease for polyatomic molecules. Furthermore, data for the two coefficients collapse onto a master curve when plotted against a translational length ratio between the liquid and vapor phase. © 2011 American Institute of Physics.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

The most feasible way to disperse particles in a bulk material or control their packing at a substrate is through fluidization in a carrier that can be processed with well-known techniques such as spin, drip and spray coating, fiber drawing, and casting. The next stage in the processing is often solidification involving drying by solvent evaporation. While there has been significant progress in the past few years in developing discrete element numerical methods to model dense nanoparticle dispersion/suspension rheology which properly treat the hydrodynamic interactions of the solvent, these methods cannot at present account for the volume reduction of the suspension due to solvent evaporation. As part of LDRD project FY-101285 we have developed and implemented methods in the current suite of discrete element methods to remove solvent particles and volume, and hence solvent mass from the liquid/vapor interface of a suspension to account for volume reduction (solvent drying) effects. To validate the methods large scale molecular dynamics simulations have been carried out to follow the evaporation process at the microscopic scale.

In this presentation we examine the accuracy and performance of a suite of discrete-element-modeling approaches to predicting equilibrium and dynamic rheological properties of polystyrene suspensions. What distinguishes each approach presented is the methodology of handling the solvent hydrodynamics. Specifically, we compare stochastic rotation dynamics (SRD), fast lubrication dynamics (FLD) and dissipative particle dynamics (DPD). Method-to-method comparisons are made as well as comparisons with experimental data. Quantities examined are equilibrium structure properties (e.g. pair-distribution function), equilibrium dynamic properties (e.g. short- and long-time diffusivities), and dynamic response (e.g. steady shear viscosity). In all approaches we deploy the DLVO potential for colloid-colloid interactions. Comparisons are made over a range of volume fractions and salt concentrations. Our results reveal the utility of such methods for long-time diffusivity prediction can be dubious in certain ranges of volume fraction, and other discoveries regarding the best formulation to use in predicting rheological response.

LAMMPS is a classical molecular dynamics code, and an acronym for Large-scale Atomic/Molecular Massively Parallel Simulator. LAMMPS has potentials for soft materials (biomolecules, polymers) and solid-state materials (metals, semiconductors) and coarse-grained or mesoscopic systems. It can be used to model atoms or, more generically, as a parallel particle simulator at the atomic, meso, or continuum scale. LAMMPS runs on single processors or in parallel using message-passing techniques and a spatial-decomposition of the simulation domain. The code is designed to be easy to modify or extend with new functionality.

Abstract not provided.

The kinetic Monte Carlo method and its variants are powerful tools for modeling materials at the mesoscale, meaning at length and time scales in between the atomic and continuum. We have completed a 3 year LDRD project with the goal of developing a parallel kinetic Monte Carlo capability and applying it to materials modeling problems of interest to Sandia. In this report we give an overview of the methods and algorithms developed, and describe our new open-source code called SPPARKS, for Stochastic Parallel PARticle Kinetic Simulator. We also highlight the development of several Monte Carlo models in SPPARKS for specific materials modeling applications, including grain growth, bubble formation, diffusion in nanoporous materials, defect formation in erbium hydrides, and surface growth and evolution.

Abstract not provided.

Abstract not provided.

Peridynamics (PD) is a continuum theory that employs a nonlocal model to describe material properties. In this context, nonlocal means that continuum points separated by a finite distance may exert force upon each other. A meshless method results when PD is discretized with material behavior approximated as a collection of interacting particles. This paper describes how PD can be implemented within a molecular dynamics (MD) framework, and provides details of an efficient implementation. This adds a computational mechanics capability to an MD code, enabling simulations at mesoscopic or even macroscopic length and time scales. © 2008 Elsevier B.V.

Abstract not provided.

Nanoparticles are now more than ever being used to tailor materials function and performance in differentiating technologies because of their profound effect on thermo-physical, mechanical and optical properties. The most feasible way to disperse particles in a bulk material or control their packing at a substrate is through fluidization in a carrier, followed by solidification through solvent evaporation/drying/curing/sintering. Unfortunately processing particles as concentrated, fluidized suspensions into useful products remains an art largely because the effect of particle shape and volume fraction on fluidic properties and suspension stability remains unexplored in a regime where particle-particle interaction mechanics is prevalent. To achieve a stronger scientific understanding of the factors that control nanoparticle dispersion and rheology we have developed a multiscale modeling approach to bridge scales between atomistic and molecular-level forces active in dense nanoparticle suspensions. At the largest length scale, two 'coarse-grained' numerical techniques have been developed and implemented to provide for high-fidelity numerical simulations of the rheological response and dispersion characteristics typical in a processing flow. The first is a coupled Navier-Stokes/discrete element method in which the background solvent is treated by finite element methods. The second is a particle based method known as stochastic rotational dynamics. These two methods provide a new capability representing a 'bridge' between the molecular scale and the engineering scale, allowing the study of fluid-nanoparticle systems over a wide range of length and timescales as well as particle concentrations. To validate these new methodologies, multi-million atoms simulations explicitly including the solvent have been carried out. These simulations have been vital in establishing the necessary 'subgrid' models for accurate prediction at a larger scale and refining the two coarse-grained methodologies.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Peridynamics is a nonlocal formulation of continuum mechanics. The discrete peridynamic model has the same computational structure as a molecular dynamic model. This document details the implementation of a discrete peridynamic model within the LAMMPS molecular dynamic code. This document provides a brief overview of the peridynamic model of a continuum, then discusses how the peridynamic model is discretized, and overviews the LAMMPS implementation. A nontrivial example problem is also included.

Abstract not provided.

Abstract not provided.

The RandomAccess benchmark as defined by the High Performance Computing Challenge (HPCC) tests the speed at which a machine can update the elements of a table spread across global system memory, as measured in billions (giga) of updates per second (GUPS). The parallel implementation provided by HPCC typically performs poorly on distributed-memory machines, due to updates requiring numerous small point-to-point messages between processors. We present an alternative algorithm which treats the collection of P processors as a hypercube, aggregating data so that larger messages are sent, and routing individual datums through dimensions of the hypercube to their destination processor. The algorithm's computation (the GUP count) scales linearly with P while its communication overhead scales as log2(P), thus enabling better performance on large numbers of processors. The new algorithm achieves a GUPS rate of 19.98 on 8192 processors of Sandia's Red Storm machine, compared to 1.02 for the HPCC-provided algorithm on 10350 processors. We also illustrate how GUPS performance varies with the benchmark's specification of its "look-ahead" parameter. As expected, parallel performance degrades for small look-ahead values, and improves dramatically for large values. © 2006 IEEE.

We have enhanced our parallel molecular dynamics (MD) simulation software LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator, lammps.sandia.gov) to include many new features for accelerated simulation including articulated rigid body dynamics via coupling to the Rensselaer Polytechnic Institute code POEMS (Parallelizable Open-source Efficient Multibody Software). We use new features of the LAMMPS software package to investigate rhodopsin photoisomerization, and water model surface tension and capillary waves at the vapor-liquid interface. Finally, we motivate the recipes of MD for practitioners and researchers in numerical analysis and computational mechanics.

Abstract not provided.

Chemically Induced Surface Evolution with Level-Sets--ChISELS--is a parallel code for modeling 2D and 3D material depositions and etches at feature scales on patterned wafers at low pressures. Designed for efficient use on a variety of computer architectures ranging from single-processor workstations to advanced massively parallel computers running MPI, ChISELS is a platform on which to build and improve upon previous feature-scale modeling tools while taking advantage of the most recent advances in load balancing and scalable solution algorithms. Evolving interfaces are represented using the level-set method and the evolution equations time integrated using a Semi-Lagrangian approach [1]. The computational meshes used are quad-trees (2D) and oct-trees (3D), constructed such that grid refinement is localized to regions near the surface interfaces. As the interface evolves, the mesh is dynamically reconstructed as needed for the grid to remain fine only around the interface. For parallel computation, a domain decomposition scheme with dynamic load balancing is used to distribute the computational work across processors. A ballistic transport model is employed to solve for the fluxes incident on each of the surface elements. Surface chemistry is computed by either coupling to the CHEMKIN software [2] or by providing user defined subroutines. This report describes the theoretical underpinnings, methods, and practical use instruction of the ChISELS 1.0 computer code.

Abstract not provided.

Abstract not provided.

Understanding the properties and behavior of biomembranes is fundamental to many biological processes and technologies. Microdomains in biomembranes or ''lipid rafts'' are now known to be an integral part of cell signaling, vesicle formation, fusion processes, protein trafficking, and viral and toxin infection processes. Understanding how microdomains form, how they depend on membrane constituents, and how they act not only has biological implications, but also will impact Sandia's effort in development of membranes that structurally adapt to their environment in a controlled manner. To provide such understanding, we created physically-based models of biomembranes. Molecular dynamics (MD) simulations and classical density functional theory (DFT) calculations using these models were applied to phenomena such as microdomain formation, membrane fusion, pattern formation, and protein insertion. Because lipid dynamics and self-organization in membranes occur on length and time scales beyond atomistic MD, we used coarse-grained models of double tail lipid molecules that spontaneously self-assemble into bilayers. DFT provided equilibrium information on membrane structure. Experimental work was performed to further help elucidate the fundamental membrane organization principles.

Abstract not provided.

Molecular dynamics calculations are performed to study the effect of deformation sequence and history on the inelastic behavior of copper interfaces on the nanoscale. An asymmetric 45 deg tilt bicrystal interface is examined, representing an idealized high-angle grain boundary interface. The interface model is subjected to three different deformation paths: tension then shear, shear then tension, and combined proportional tension and shear. Analysis shows that path-history dependent material behavior is confined within a finite layer of deformation around the bicrystal interface. The relationships between length scale and interface properties, such as the thickness of the path-history dependent layer and the interface strength, are discussed in detail.