-
Enabling particle applications for exascale computing platforms
Authors:
Susan M Mniszewski,
James Belak,
Jean-Luc Fattebert,
Christian FA Negre,
Stuart R Slattery,
Adetokunbo A Adedoyin,
Robert F Bird,
Choongseok Chang,
Guangye Chen,
Stephane Ethier,
Shane Fogerty,
Salman Habib,
Christoph Junghans,
Damien Lebrun-Grandie,
Jamaludin Mohd-Yusof,
Stan G Moore,
Daniel Osei-Kuffuor,
Steven J Plimpton,
Adrian Pope,
Samuel Temple Reeve,
Lee Ricketson,
Aaron Scheinberg,
Amil Y Sharma,
Michael E Wall
Abstract:
The Exascale Computing Project (ECP) is invested in co-design to assure that key applications are ready for exascale computing. Within ECP, the Co-design Center for Particle Applications (CoPA) is addressing challenges faced by particle-based applications across four sub-motifs: short-range particle-particle interactions (e.g., those which often dominate molecular dynamics (MD) and smoothed partic…
▽ More
The Exascale Computing Project (ECP) is invested in co-design to assure that key applications are ready for exascale computing. Within ECP, the Co-design Center for Particle Applications (CoPA) is addressing challenges faced by particle-based applications across four sub-motifs: short-range particle-particle interactions (e.g., those which often dominate molecular dynamics (MD) and smoothed particle hydrodynamics (SPH) methods), long-range particle-particle interactions (e.g., electrostatic MD and gravitational N-body), particle-in-cell (PIC) methods, and linear-scaling electronic structure and quantum molecular dynamics (QMD) algorithms. Our crosscutting co-designed technologies fall into two categories: proxy applications (or apps) and libraries. Proxy apps are vehicles used to evaluate the viability of incorporating various types of algorithms, data structures, and architecture-specific optimizations and the associated trade-offs; examples include ExaMiniMD, CabanaMD, CabanaPIC, and ExaSP2. Libraries are modular instantiations that multiple applications can utilize or be built upon; CoPA has developed the Cabana particle library, PROGRESS/BML libraries for QMD, and the SWFFT and fftMPI parallel FFT libraries. Success is measured by identifiable lessons learned that are translated either directly into parent production application codes or into libraries, with demonstrated performance and/or productivity improvement. The libraries and their use in CoPA's ECP application partner codes are also addressed.
△ Less
Submitted 19 September, 2021;
originally announced September 2021.
-
Granular packings with sliding, rolling and twisting friction
Authors:
A. P. Santos,
Dan S. Bolintineanu,
Gary S. Grest,
Jeremy B. Lechman,
Steven J. Plimpton,
Ishan Srivastava,
Leonardo E. Silbert
Abstract:
Intuition tells us that a rolling or spinning sphere will eventually stop due to the presence of friction and other dissipative interactions. The resistance to rolling and spinning/twisting torque that stops a sphere also changes the microstructure of a granular packing of frictional spheres by increasing the number of constraints on the degrees of freedom of motion. We perform discrete element mo…
▽ More
Intuition tells us that a rolling or spinning sphere will eventually stop due to the presence of friction and other dissipative interactions. The resistance to rolling and spinning/twisting torque that stops a sphere also changes the microstructure of a granular packing of frictional spheres by increasing the number of constraints on the degrees of freedom of motion. We perform discrete element modeling simulations to construct sphere packings implementing a range of frictional constraints under a pressure-controlled protocol. Mechanically stable packings are achievable at volume fractions and average coordination numbers as low as 0.53 and 2.5, respectively, when the particles experience high resistance to sliding, rolling and twisting. Only when the particle model includes rolling and twisting friction, were experimental volume fractions reproduced.
△ Less
Submitted 21 July, 2020;
originally announced July 2020.
-
Effect of Shape and Friction on the Packing and Flow of Granular Materials
Authors:
K. Michael Salerno,
Dan S. Bolintineanu,
Gary S. Grest,
Jeremy B. Lechman,
Steven J. Plimpton,
Ishan Srivastava,
Leonardo E. Silbert
Abstract:
The packing and flow of aspherical frictional particles are studied using discrete element simulations. Particles are superballs with shape $|x|^{s}+|y|^{s}+|z|^{s} = 1$ that varies from sphere ($s=2$) to cube ($s=\infty$), constructed with an overlapping-sphere model. Both packing fraction, $φ$, and coordination number, $z$, decrease monotonically with microscopic friction $μ$, for all shapes. Ho…
▽ More
The packing and flow of aspherical frictional particles are studied using discrete element simulations. Particles are superballs with shape $|x|^{s}+|y|^{s}+|z|^{s} = 1$ that varies from sphere ($s=2$) to cube ($s=\infty$), constructed with an overlapping-sphere model. Both packing fraction, $φ$, and coordination number, $z$, decrease monotonically with microscopic friction $μ$, for all shapes. However, this decrease is more dramatic for larger $s$ due to a reduction in the fraction of face-face contacts with increasing friction. For flowing grains, the dynamic friction $\tildeμ$ - the ratio of shear to normal stresses - depends on shape, microscopic friction and inertial number $I.$ For all shapes, $\tildeμ$ grows from its quasi-static value $\tildeμ_0$ as $(\tildeμ-\tildeμ_0) = dI^α,$ with different universal behavior for frictional and frictionless shapes. For frictionless shapes the exponent $α\approx 0.5$ and prefactor $d \approx 5\tildeμ_0$ while for frictional shapes $α\approx 1$ and $d$ varies only slightly. The results highlight that the flow exponents are universal and are consistent for all the shapes simulated here.
△ Less
Submitted 31 October, 2018;
originally announced October 2018.
-
Massively parallel symplectic algorithm for coupled magnetic spin dynamics and molecular dynamics
Authors:
J. Tranchida,
S. J. Plimpton,
P. Thibaudeau,
A. P. Thompson
Abstract:
A parallel implementation of coupled spin-lattice dynamics in the LAMMPS molecular dynamics package is presented. The equations of motion for both spin only and coupled spin-lattice dynamics are first reviewed, including a detailed account of how magneto-mechanical potentials can be used to perform a proper coupling between spin and lattice degrees of freedom. A symplectic numerical integration al…
▽ More
A parallel implementation of coupled spin-lattice dynamics in the LAMMPS molecular dynamics package is presented. The equations of motion for both spin only and coupled spin-lattice dynamics are first reviewed, including a detailed account of how magneto-mechanical potentials can be used to perform a proper coupling between spin and lattice degrees of freedom. A symplectic numerical integration algorithm is then presented which combines the Suzuki-Trotter decomposition for non-commuting variables and conserves the geometric properties of the equations of motion. The numerical accuracy of the serial implementation was assessed by verifying that it conserves the total energy and the norm of the total magnetization up to second order in the timestep size. Finally, a very general parallel algorithm is proposed that allows large spin-lattice systems to be efficiently simulated on large numbers of processors without degrading its mathematical accuracy. Its correctness as well as scaling efficiency were tested for realistic coupled spin-lattice systems, confirming that the new parallel algorithm is both accurate and efficient.
△ Less
Submitted 30 January, 2018;
originally announced January 2018.
-
Multiscale Co-Design Analysis of Energy, Latency, Area, and Accuracy of a ReRAM Analog Neural Training Accelerator
Authors:
Matthew J. Marinella,
Sapan Agarwal,
Alexander Hsia,
Isaac Richter,
Robin Jacobs-Gedrim,
John Niroula,
Steven J. Plimpton,
Engin Ipek,
Conrad D. James
Abstract:
Neural networks are an increasingly attractive algorithm for natural language processing and pattern recognition. Deep networks with >50M parameters are made possible by modern GPU clusters operating at <50 pJ per op and more recently, production accelerators capable of <5pJ per operation at the board level. However, with the slowing of CMOS scaling, new paradigms will be required to achieve the n…
▽ More
Neural networks are an increasingly attractive algorithm for natural language processing and pattern recognition. Deep networks with >50M parameters are made possible by modern GPU clusters operating at <50 pJ per op and more recently, production accelerators capable of <5pJ per operation at the board level. However, with the slowing of CMOS scaling, new paradigms will be required to achieve the next several orders of magnitude in performance per watt gains. Using an analog resistive memory (ReRAM) crossbar to perform key matrix operations in an accelerator is an attractive option. This work presents a detailed design using a state of the art 14/16 nm PDK for of an analog crossbar circuit block designed to process three key kernels required in training and inference of neural networks. A detailed circuit and device-level analysis of energy, latency, area, and accuracy are given and compared to relevant designs using standard digital ReRAM and SRAM operations. It is shown that the analog accelerator has a 270x energy and 540x latency advantage over a similar block utilizing only digital ReRAM and takes only 11 fJ per multiply and accumulate (MAC). Compared to an SRAM based accelerator, the energy is 430X better and latency is 34X better. Although training accuracy is degraded in the analog accelerator, several options to improve this are presented. The possible gains over a similar digital-only version of this accelerator block suggest that continued optimization of analog resistive memories is valuable. This detailed circuit and device analysis of a training accelerator may serve as a foundation for further architecture-level studies.
△ Less
Submitted 16 February, 2018; v1 submitted 31 July, 2017;
originally announced July 2017.
-
Evaporation of Lennard-Jones Fluids
Authors:
Shengfeng Cheng,
Jeremy B. Lechman,
Steven J. Plimpton,
Gary S. Grest
Abstract:
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 r…
▽ More
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 $T/T_c$, where $T_c$ 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.
△ Less
Submitted 9 May, 2011; v1 submitted 27 January, 2011;
originally announced January 2011.
-
Novel statistical ensemble analysis for simulating extrinsic noise-driven response in NF-κB signaling network
Authors:
Jaewook Joo,
Steven J. Plimpton,
Jean-Loup Faulon
Abstract:
Cellular responses in the single cells are known to be highly heterogeneous and individualistic due to the strong influence by extrinsic and intrinsic noise. Here, we are concerned about how to model the extrinsic noise-induced heterogeneous response in the single cells under the constraints of experimentally obtained population-averaged response, but without much detailed kinetic information. We…
▽ More
Cellular responses in the single cells are known to be highly heterogeneous and individualistic due to the strong influence by extrinsic and intrinsic noise. Here, we are concerned about how to model the extrinsic noise-induced heterogeneous response in the single cells under the constraints of experimentally obtained population-averaged response, but without much detailed kinetic information. We propose a novel statistical ensemble scheme where extrinsic noise is regarded as fluctuations in the values of kinetic parameters and such fluctuations are modeled by randomly sampling the kinetic rate constants from a uniform distribution. We consider a large number of signaling system replicates, each of which has the same network topology, but a uniquely different set of kinetic rate constants. A protein dynamic response from each replicate should represent the dynamics in a single cell and the statistical ensemble average should be regarded as a population-level response averaged over a population of the cells. We devise an optimization algorithm to find the correct uniform distribution of the network parameters, which produces the correct statistical distribution of the response whose ensemble average and distribution agree well with the population-level experimental data and the experimentally observed heterogeneity. We apply this statistical ensemble analysis to a NF-κB signaling system and (1) predict the distributions of the heterogeneous NF-κB (either oscillatory or non-oscillatory) dynamic patterns and of the dynamic features (e.g., period), (2) predict that both the distribution and the statistical ensemble average of the NF-κB dynamic response depends sensitively on the dosage of stimulant, and lastly (3) demonstrate the sigmoidally shaped dose-response from the statistical ensemble average and the individual replicates.
△ Less
Submitted 5 October, 2010;
originally announced October 2010.
-
Noise-induced oscillatory shuttling of NF-κB in a two compartment IKK-NF-κB-IκB-A20 signaling model
Authors:
Jaewook Joo,
Steven J. Plimpton,
Jean-Loup Faulon
Abstract:
NF-κB is a pleiotropic protein whose nucleo-cytoplasmic trafficking is tightly regulated by multiple negative feedback loops embedded in the NF-κB signaling network and contributes to diverse gene expression profiles important in immune cell differentiation, cell apoptosis, and innate immunity. The intracellular signaling processes and their control mechanisms, however, are susceptible to both ext…
▽ More
NF-κB is a pleiotropic protein whose nucleo-cytoplasmic trafficking is tightly regulated by multiple negative feedback loops embedded in the NF-κB signaling network and contributes to diverse gene expression profiles important in immune cell differentiation, cell apoptosis, and innate immunity. The intracellular signaling processes and their control mechanisms, however, are susceptible to both extrinsic and intrinsic noise. In this article, we present numerical evidence for a universal dynamic behavior of NF-κB, namely oscillatory nucleo-cytoplasmic shuttling, due to the fundamentally stochastic nature of the NF-κB signaling network. We simulated the effect of extrinsic noise with a deterministic ODE model, using a statistical ensemble approach, generating many copies of the signaling network with different kinetic rates sampled from a biologically feasible parameter space. We modeled the effect of intrinsic noise by simulating the same networks stochastically using the Gillespie algorithm. The results demonstrate that extrinsic noise diversifies the shuttling patterns of NF-κB response, whereas intrinsic noise induces oscillatory behavior in many of the otherwise non-oscillatory patterns. We identify two key model parameters which significantly affect the NF-κB dynamic response and deduce a two-dimensional phase-diagram of the NF-κB response as a function of these parameters. We conclude that if single-cell experiments are performed, a rich variety of NF-κB response will be observed, even if population-level experiments, which average response over large numbers of cells, do not evidence oscillatory behavior.
△ Less
Submitted 5 October, 2010;
originally announced October 2010.
-
Equilibration of Long Chain Polymer Melts in Computer Simulations
Authors:
Rolf Auhl,
Ralf Everaers,
Gary S. Grest,
Kurt Kremer,
Steven J. Plimpton
Abstract:
Several methods for preparing well equilibrated melts of long chains polymers are studied. We show that the standard method in which one starts with an ensemble of chains with the correct end-to-end distance arranged randomly in the simulation cell and introduces the excluded volume rapidly, leads to deformation on short length scales. This deformation is strongest for long chains and relaxes on…
▽ More
Several methods for preparing well equilibrated melts of long chains polymers are studied. We show that the standard method in which one starts with an ensemble of chains with the correct end-to-end distance arranged randomly in the simulation cell and introduces the excluded volume rapidly, leads to deformation on short length scales. This deformation is strongest for long chains and relaxes only after the chains have moved their own size. Two methods are shown to overcome this local deformation of the chains. One method is to first pre-pack the Gaussian chains, which reduces the density fluctuations in the system, followed by a gradual introduction of the excluded volume. The second method is a double-pivot algorithm in which new bonds are formed across a pair of chains, creating two new chains each substantially different from the original. We demonstrate the effectiveness of these methods for a linear bead spring polymer model with both zero and nonzero bending stiffness, however the methods are applicable to more complex architectures such as branched and star polymer.
△ Less
Submitted 2 June, 2003;
originally announced June 2003.
-
Discrete element simulations of stress distributions in silos: crossover from two to three dimensions
Authors:
James W. Landry,
Gary S. Grest,
Stephen J. Plimpton
Abstract:
The transition from two-dimensional (2D) to three-dimensional (3D) granular packings is studied using large-scale discrete element computer simulations. We focus on vertical stress profiles and examine how they change with dimensionality from 2D to 3D. We compare results for packings in 2D, quasi-2D packings between flat plates, and 3D packings. Analysis of these packings suggests that the Janss…
▽ More
The transition from two-dimensional (2D) to three-dimensional (3D) granular packings is studied using large-scale discrete element computer simulations. We focus on vertical stress profiles and examine how they change with dimensionality from 2D to 3D. We compare results for packings in 2D, quasi-2D packings between flat plates, and 3D packings. Analysis of these packings suggests that the Janssen theory does not fully describe these packings, especially at the top of the piles, where a hydrostatic-like region of vertical stress is visible in all cases. We find that the interior of the packing is far from incipient failure, while in general, the forces at the walls are close to incipient failure.
△ Less
Submitted 4 November, 2003; v1 submitted 6 February, 2003;
originally announced February 2003.
-
Confined granular packings: structure, stress, and forces
Authors:
J. W. Landry,
G. S. Grest,
L. E. Silbert,
S. J. Plimpton
Abstract:
The structure and stresses of static granular packs in cylindrical containers are studied using large-scale discrete element molecular dynamics simulations in three dimensions. We generate packings by both pouring and sedimentation and examine how the final state depends on the method of construction. The vertical stress becomes depth-independent for deep piles and we compare these stress depth-…
▽ More
The structure and stresses of static granular packs in cylindrical containers are studied using large-scale discrete element molecular dynamics simulations in three dimensions. We generate packings by both pouring and sedimentation and examine how the final state depends on the method of construction. The vertical stress becomes depth-independent for deep piles and we compare these stress depth-profiles to the classical Janssen theory. The majority of the tangential forces for particle-wall contacts are found to be close to the Coulomb failure criterion, in agreement with the theory of Janssen, while particle-particle contacts in the bulk are far from the Coulomb criterion. In addition, we show that a linear hydrostatic-like region at the top of the packings unexplained by the Janssen theory arises because most of the particle-wall tangential forces in this region are far from the Coulomb yield criterion. The distributions of particle-particle and particle-wall contact forces $P(f)$ exhibit exponential-like decay at large forces in agreement with previous studies.
△ Less
Submitted 31 January, 2003; v1 submitted 11 November, 2002;
originally announced November 2002.
-
Granular flow down an inclined plane: Bagnold scaling and rheology
Authors:
L. E. Silbert,
D. Ertas,
G. S. Grest,
T. C. Halsey,
D. Levine,
S. J. Plimpton
Abstract:
We have performed a systematic, large-scale simulation study of granular media in two- and three-dimensions, investigating the rheology of cohesionless granular particles in inclined plane geometries, i.e., chute flows. We find that over a wide range of parameter space of interaction coefficients and inclination angles, a steady state flow regime exists in which the energy input from gravity bal…
▽ More
We have performed a systematic, large-scale simulation study of granular media in two- and three-dimensions, investigating the rheology of cohesionless granular particles in inclined plane geometries, i.e., chute flows. We find that over a wide range of parameter space of interaction coefficients and inclination angles, a steady state flow regime exists in which the energy input from gravity balances that dissipated from friction and inelastic collisions. In this regime, the bulk packing fraction (away from the top free surface and the bottom plate boundary) remains constant as a function of depth z, of the pile. The velocity profile in the direction of flow v(z) scales with height of the pile H, according to v(z) proportional to H^a, with a=1.52. However, the behavior of the normal stresses indicates that existing simple theories of granular flow do not capture all of the features evidenced in the simulations.
△ Less
Submitted 3 May, 2001;
originally announced May 2001.
-
Spatial Correlations of Mobility and Immobility in a Glassforming Lennard-Jones Liquid
Authors:
C. Donati,
S. C. Glotzer,
P. H. Poole,
W. Kob,
S. J. Plimpton
Abstract:
Using extensive molecular dynamics simulations of an equilibrium, glass-forming Lennard-Jones mixture, we characterize in detail the local atomic motions. We show that spatial correlations exist among particles undergoing extremely large (``mobile'') or extremely small (``immobile'') displacements over a suitably chosen time interval. The immobile particles form the cores of relatively compact c…
▽ More
Using extensive molecular dynamics simulations of an equilibrium, glass-forming Lennard-Jones mixture, we characterize in detail the local atomic motions. We show that spatial correlations exist among particles undergoing extremely large (``mobile'') or extremely small (``immobile'') displacements over a suitably chosen time interval. The immobile particles form the cores of relatively compact clusters, while the mobile particles move cooperatively and form quasi-one-dimensional, string-like clusters. The strength and length scale of the correlations between mobile particles are found to grow strongly with decreasing temperature, and the mean cluster size appears to diverge at the mode-coupling critical temperature. We show that these correlations in the particle displacements are related to equilibrium fluctuations in the local potential energy and local composition.
△ Less
Submitted 6 October, 1998;
originally announced October 1998.
-
String-like Clusters and Cooperative Motion in a Model Glass-Forming Liquid
Authors:
Claudio Donati,
Jack F. Douglas,
Walter Kob,
Steven J. Plimpton,
Peter H. Poole,
Sharon C. Glotzer
Abstract:
A large-scale molecular dynamics simulation is performed on a glass-forming Lennard-Jones mixture to determine the nature of dynamical heterogeneities which arise in this model fragile liquid. We observe that the most mobile particles exhibit a cooperative motion in the form of string-like paths (``strings'') whose mean length and radius of gyration increase as the liquid is cooled. The length d…
▽ More
A large-scale molecular dynamics simulation is performed on a glass-forming Lennard-Jones mixture to determine the nature of dynamical heterogeneities which arise in this model fragile liquid. We observe that the most mobile particles exhibit a cooperative motion in the form of string-like paths (``strings'') whose mean length and radius of gyration increase as the liquid is cooled. The length distribution of the strings is found to be similar to that expected for the equilibrium polymerization of linear polymer chains.
△ Less
Submitted 27 June, 1997;
originally announced June 1997.
-
Dynamical heterogeneities in a supercooled Lennard-Jones liquid
Authors:
Walter Kob,
Claudio Donati,
Steven J. Plimpton,
Peter H. Poole,
Sharon C. Glotzer
Abstract:
We present the results of a large scale molecular dynamics computer simulation study in which we investigate whether a supercooled Lennard-Jones liquid exhibits dynamical heterogeneities. We evaluate the non-Gaussian parameter for the self part of the van Hove correlation function and use it to identify ``mobile'' particles. We find that these particles form clusters whose size grows with decrea…
▽ More
We present the results of a large scale molecular dynamics computer simulation study in which we investigate whether a supercooled Lennard-Jones liquid exhibits dynamical heterogeneities. We evaluate the non-Gaussian parameter for the self part of the van Hove correlation function and use it to identify ``mobile'' particles. We find that these particles form clusters whose size grows with decreasing temperature. We also find that the relaxation time of the mobile particles is significantly shorter than that of the bulk, and that this difference increases with decreasing temperature.
△ Less
Submitted 8 June, 1997;
originally announced June 1997.
