Search NASA⌕ Search

SEARCH · Search NASA

Results for “distributed parallelization”

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 937 records · Page 52

Evaluation of turbulent magnetic energy spectra in the three-dimensional wave vector domain in the solar wind

Using four-point measurements of the CLUSTER spacecraft, the energy distribution of magnetic field fluctuations in the solar wind is determined directly in the three-dimensional wave vector domain in the range 3 x 10{sup -4} rad/km < k < 3 x 10{sup -3} rad/km. The analysis method takes account of a regular tetrahedron configuration of CLUSTER and the Doppler effect. The energy distribution in the flow rest frame is anisotropic, characterized by two distinct extended structures perpendicular to the mean magnetic field and furthermore perpendicular to the flow direction. The three-dimensional distribution is averaged around the direction of the mean magnetic field direction, and then is further reduced to one-dimensional distributions in the wave number domain parallel and perpendicular to the mean magnetic field. The one-dimensional energy spectra are characterized by the power law with the index -5/3 and furthermore very close energy density between parallel and perpendicular directions to the mean magnetic field at the same wave numbers. Though the distributions and the spectra are not covered in a wide range of wave vectors, our measurements suggest that the solar wind fluctuation is anisotropic in the three-dimensional wave vector space. It is, however, rather isotropic when reduced into the parallel and perpendicular wave vector geometries due to the second anisotropy imposed by the flow direction.

58 GEOSCIENCES↗

The constraint of plasma power balance on runaway avoidance

Abstract In a post-thermal-quench plasma, mitigated or unmitigated, the plasma power balance is mostly between collisional or Ohmic heating and plasma radiative cooling. In a plasma of atomic mixture { n α } with α labeling the atomic species, the power balance sets the plasma temperature, ion charge state distribution { n α i } with i the charge number, and through the electron temperature T e and ion charge state distribution { n α i } , the parallel electric field E ∥ . Since the threshold electric field for runaway avalanche growth E a v is also set by the atomic mixture, ion charge state distribution and its derived quantity, the electron density n e , the plasma power balance between Ohmic heating and radiative cooling imposes a stringent constraint on the plasma regime for avoiding and minimizing runaways when a fusion-grade tokamak plasma is rapidly terminated.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

SIMULATION OF ENERGETIC PARTICLE TRANSPORT AND ACCELERATION AT SHOCK WAVES IN A FOCUSED TRANSPORT MODEL: IMPLICATIONS FOR MIXED SOLAR PARTICLE EVENTS

We use numerical solutions of the focused transport equation obtained by an implicit stochastic differential equation scheme to study the evolution of the pitch-angle dependent distribution function of protons in the vicinity of shock waves. For a planar stationary parallel shock, the effects of anisotropic distribution functions, pitch-angle dependent spatial diffusion, and first-order Fermi acceleration at the shock are examined, including the timescales on which the energy spectrum approaches the predictions of diffusive shock acceleration theory. We then consider the case that a flare-accelerated population of ions is released close to the Sun simultaneously with a traveling interplanetary shock for which we assume a simplified geometry. We investigate the consequences of adiabatic focusing in the diverging magnetic field on the particle transport at the shock, and of the competing effects of acceleration at the shock and adiabatic energy losses in the expanding solar wind. We analyze the resulting intensities, anisotropies, and energy spectra as a function of time and find that our simulations can naturally reproduce the morphologies of so-called mixed particle events in which sometimes the prompt and sometimes the shock component is more prominent, by assuming parameter values which are typically observed for scattering mean free paths of ions in the inner heliosphere and energy spectra of the flare particles which are injected simultaneously with the release of the shock.

79 ASTRONOMY AND ASTROPHYSICS↗

Proton beam generation of whistler waves in the earth's foreshock

It is shown that proton beams, often observed upstream of the earth's bow shock and associated with the generation of low-frequency hydromagnetic fluctuations, are also capable of generating whistler waves. The waves can be excited by an instability driven by two-temperature streaming Maxwellian proton distributions which have T (perpendicular)/T(parallel) much greater than 1. It can also be excited by gyrating proton beam distributions. These distributions generate whistler waves with frequencies ranging from 10 to 100 times the proton cyclotron frequency (in the solar wind reference frame) and provide another mechanism for generating the '1-Hz' waves often seen in the earth's foreshock.

Wong, H. K.↗

Distributed neural control of a hexapod walking vehicle

There has been a long standing interest in the design of controllers for multilegged vehicles. The approach is to apply distributed control to this problem, rather than using parallel computing of a centralized algorithm. Researchers describe a distributed neural network controller for hexapod locomotion which is based on the neural control of locomotion in insects. The model considers the simplified kinematics with two degrees of freedom per leg, but the model includes the static stability constraint. Through simulation, it is demonstrated that this controller can generate a continuous range of statically stable gaits at different speeds by varying a single control parameter. In addition, the controller is extremely robust, and can continue the function even after several of its elements have been disabled. Researchers are building a small hexapod robot whose locomotion will be controlled by this network. Researchers intend to extend their model to the dynamic control of legs with more than two degrees of freedom by using data on the control of multisegmented insect legs. Another immediate application of this neural control approach is also exhibited in biology: the escape reflex. Advanced robots are being equipped with tactile sensing and machine vision so that the sensory inputs to the robot controller are vast and complex. Neural networks are ideal for a lower level safety reflex controller because of their extremely fast response time. The combination of robotics, computer modeling, and neurobiology has been remarkably fruitful, and is likely to lead to deeper insights into the problems of real time sensorimotor control.

Beer, R. D.↗

Devastator Parallel Discrete Event Simulation Runtime (Devastator) v1.0

The Devastator runtime is a modern C++ implementation of optimistic parallel discrete event simulation methods. Devastator allows simulation application code to productively specify their component and event functionality with C++14 constructs. It utilizes GASNet-EX for distributed memory communication and includes parallel performance optimizations such as light-weight thread message queues and asynchronous GVT. Furthermore, it supports efficient event broadcasts and pause-rewind-resume functionality to support periodic load balancing and outer loop optimization algorithms.

Chan, Cy↗

Execution time supports for adaptive scientific algorithms on distributed memory machines

Optimizations are considered that are required for efficient execution of code segments that consists of loops over distributed data structures. The PARTI (Parallel Automated Runtime Toolkit at ICASE) execution time primitives are designed to carry out these optimizations and can be used to implement a wide range of scientific algorithms on distributed memory machines. These primitives allow the user to control array mappings in a way that gives an appearance of shared memory. Computations can be based on a global index set. Primitives are used to carry out gather and scatter operations on distributed arrays. Communications patterns are derived at runtime, and the appropriate send and receive messages are automatically generated.

Berryman, Harry↗

Execution time support for scientific programs on distributed memory machines

Optimizations are considered that are required for efficient execution of code segments that consists of loops over distributed data structures. The PARTI (Parallel Automated Runtime Toolkit at ICASE) execution time primitives are designed to carry out these optimizations and can be used to implement a wide range of scientific algorithms on distributed memory machines. These primitives allow the user to control array mappings in a way that gives an appearance of shared memory. Computations can be based on a global index set. Primitives are used to carry out gather and scatter operations on distributed arrays. Communications patterns are derived at runtime, and the appropriate send and receive messages are automatically generated.

Berryman, Harry↗

Mapping a battlefield simulation onto message-passing parallel architectures

Perhaps the most critical problem in distributed simulation is that of mapping: without an effective mapping of workload to processors the speedup potential of parallel processing cannot be realized. Mapping a simulation onto a message-passing architecture is especially difficult when the computational workload dynamically changes as a function of time and space; this is exactly the situation faced by battlefield simulations. This paper studies an approach where the simulated battlefield domain is first partitioned into many regions of equal size; typically there are more regions than processors. The regions are then assigned to processors; a processor is responsible for performing all simulation activity associated with the regions. The assignment algorithm is quite simple and attempts to balance load by exploiting locality of workload intensity. The performance of this technique is studied on a simple battlefield simulation implemented on the Flex/32 multiprocessor. Measurements show that the proposed method achieves reasonable processor efficiencies. Furthermore, the method shows promise for use in dynamic remapping of the simulation.

Nicol, David M.↗

Design and performance of a scalable, parallel statistics toolkit.

Most statistical software packages implement a broad range of techniques but do so in an ad hoc fashion, leaving users who do not have a broad knowledge of statistics at a disadvantage since they may not understand all the implications of a given analysis or how to test the validity of results. These packages are also largely serial in nature, or target multicore architectures instead of distributed-memory systems, or provide only a small number of statistics in parallel. This paper surveys a collection of parallel implementations of statistics algorithm developed as part of a common framework over the last 3 years. The framework strategically groups modeling techniques with associated verification and validation techniques to make the underlying assumptions of the statistics more clear. Furthermore it employs a design pattern specifically targeted for distributed-memory parallelism, where architectural advances in large-scale high-performance computing have been focused. Moment-based statistics (which include descriptive, correlative, and multicorrelative statistics, principal component analysis (PCA), and k-means statistics) scale nearly linearly with the data set size and number of processes. Entropy-based statistics (which include order and contingency statistics) do not scale well when the data in question is continuous or quasi-diffuse but do scale well when the data is discrete and compact. We confirm and extend our earlier results by now establishing near-optimal scalability with up to 10,000 processes.

99 GENERAL AND MISCELLANEOUS↗

A template-based approach for parallel hexahedral two-refinement

Here, we provide a template-based approach for generating locally refined all-hex meshes. We focus specifically on refinement of initially structured grids utilizing a 2-refinement approach where uniformly refined hexes are subdivided into eight child elements. The refinement algorithm consists of identifying marked nodes that are used as the basis for a set of four simple refinement templates. The target application for 2-refinement is a parallel grid-based all-hex meshing tool for high performance computing in a distributed environment. The result is a parallel consistent locally refined mesh requiring minimal communication and where minimum mesh quality is greater than scaled Jacobian 0.3 prior to smoothing.

2-refinement↗

Wave-Particle Interactions Associated with Nongyrotropic Distribution Functions: A Hybrid Simulation Study

Nongyrotropic plasma distribution functions can be formed in regions of space where guiding center motion breaks down as a result of strongly curved and weak ambient magnetic fields. Such are the conditions near the current sheet in the Earth's middle and distant magnetotail, where observations of nongyrotropic ion distributions have been made. Here a systematic parameter study of nongyrotropic proton distributions using electromagnetic hybrid simulations is made. We model the observed nongyrotropic distributions by removing a number of arc length segments from a cold ring distribution and find significant differences with the results of simulations that initially have a gyrotropic ring distribution. Model nongyrotropic distributions with initially small perpendicular thermalization produce growing fluctuations that diffuse the ions into a stable Maxwellian-like distribution within a few proton gyro periods. The growing waves produced by nongyrotropic distributions are similar to the electromagnetic proton cyclotron waves produced by a gyrotropic proton ring distribution in that they propagate parallel to the background magnetic field and occur at frequencies on the order of the proton gyrofrequency, The maximum energy of the fluctuating magnetic field increases as the initial proton distribution is made more nongyrotropic, that is, more highly bunched in perpendicular velocity space. This increase can be as much as twice the energy produced in the gyrotropic case.

Convery, P. D.↗

Observations of Closed Structures at the Magnetopause: A Case for Multiple Reconnections

We further analyze a case of Interball LLBL crossing on the dusk flank of geomagnetosphere under southward magnetosheath magnetic field, previously categorized as an interval of highly structured LLBL. These conditions of highly structured LLBL include reconnection signatures. Observed ion velocity distributions with LLBL are quite variable. D-shaped distributions that are associated with the open reconnected flux tube are observed at the boundaries of LLBL transients and sometimes within the LLBL transients. In most cases the ion velocity distributions consist of two magnetosheath-type components with different velocities parallel to the magnetic field, or of three components one of which has nearly zero Vpar. The shapes of ion velocity distributions and their evolution with decreasing number density in LLBL indicate that most of LLBL is located on closed magnetic field lines. These observations strongly favor multiple reconnections between magnetosheath and magnetosphereric flux tubes, creating long spiral flux tube islands at the magnetopause. We report evidence for the simultaneous occurrence of magnetic reconnection at multiple points across the magnetopause, as has been proposed and found to occur in magnetopause simulations. The evidence is in the form of highly structured distributions of ions in velocity parallel to the local magnetic field direction, within the magnetopause and low latitude boundary layer region, from the Interball-Tall spacecraft. We interpret these distributions as a natural consequence of the formation of spiral magnetic flux tubes consisting of a mixture of alternating segments originating from the magnetosheath or interplanetary plasma and from the low latitude boundary layer or magnetospheric plasma. We further analyze a case of Interball LLBL crossing on the dusk flank of geomagnetosphere under southward magnetosheath magnetic field, previously categorized as an interval of highly structured LLBL. These conditions of highly structured LLBL include reconnection signatures. Observed ion velocity distributions with LLBL are quite variable. D-shaped distributions that are associated with the open reconnected flux tube are observed at the boundaries of LLBL transients and sometimes within the LLBL transients. In most cases the ion velocity distributions consist of two magnetosheath-type components with different velocities parallel to the magnetic field, or of three components one of which has nearly zero Vpar. The shapes of ion velocity distributions and their evolution with decreasing number density in LLBL indicate that most of LLBL is located on closed magnetic field lines. These observations strongly favor multiple reconnection between magnetosheath and magnetospheric flux tubes, creating long spiral flux tube islands at the magnetopause. We report evidence for the simultaneous occurrence of magnetic reconnection at multiple points across the magnetopause, as has been proposed and found to occur in magnetopause simulations. The evidence is in the form of highly structured distributions of ions in velocity parallel to the local magnetic field direction, within the magnetopause and low latitude boundary layer region, from the Interball-Tail spacecraft. We interpret these distributions as a natural consequence of the formation of spiral. We further analyze a case of Interball LLBL crossing on the dusk flank of geomagnetosphere under southward magnetosheath magnetic field, previously categorized as an interval of highly structured LLBL. These conditions of highly structured LLBL include reconnection signatures. Observed ion velocity distributions with LLBL are quite variable. D-shaped distributions that are associated with the open reconnected flux tube are observed at the boundaries of LLBL transients and sometimes within the LLBL transients. In most cases the ion velocity distributions consist of two magnetosheath-type components with different velocities parallel to the magnetic field, or of three components one of which has nearly zero Vpar.

Vaisberg, O. L.↗

Computation of absolute hydration and binding free energy with free energy perturbation distributed replica-exchange molecular dynamics.

Distributed Replica (REPDSTR) is a powerful parallelization technique enabling simulations of a group of replicas in a parallel/parallel fashion, where each replica is distributed to different nodes of a large cluster [Theor. Chem. Acc. 2003, 109, 140]. Here, we use the framework provided by REPDSTR to combine a staged free energy perturbation protocol with replica-exchange molecular dynamics (FEP/REMD). The structure of REPDSTR, which allows multiple parallel input/output (I/O), facilitates the treatment of replica-exchange to couple the N window simulations corresponding to different values of the thermodynamic coupling parameters. As a result, each of the N synchronous window simulations benefit from the sampling carried out by the N-1 others. As illustrative examples of the FEP/REMD strategy, calculations of the absolute hydration and binding free energy of small molecules were performed using the biomolecular simulation program CHARMM adapted for the IBM Blue Gene/P platform. The computations show that a FEP/REMD strategy significantly improves the sampling and accelerates the convergence of absolute free energy computations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Molecular dynamics of the water liquid-vapor interface

The results of molecular dynamics calculations on the equilibrium interface between liquid water and its vapor at 325 K are presented. For the TIP4P model of water intermolecular pair potentials, the average surface dipole density points from the vapor to the liquid. The most common orientations of water molecules have the C2 nu molecular axis roughly parallel to the interface. The distributions are quite broad and therefore compatible with the intermolecular correlations characteristic of bulk liquid water. All near-neighbor pairs in the outermost interfacial layers are hydrogen bonded according to the common definition adopted here. The orientational preferences of water molecules near a free surface differ from those near rigidly planar walls which can be interpreted in terms of patterns found in hexagonal ice 1. The mean electric field in the interfacial region is parallel to the mean polarization which indicates that attention cannot be limited to dipolar charge distributions in macroscopic descriptions of the electrical properties of this interface. The value of the surface tension obtained is 132 +/- 46 dyn/cm, significantly different from the value for experimental water of 68 dyn/cm at 325 K.

NASA Center ARC↗

Evaluating the Practicability of Multiphase CFD for BWR Fuel Assembly Analysis

Multiphase computational fluid dynamics (M-CFD) models of a boiling water reactor (BWR) fuel assembly have been generated in the commercial CFD code, STAR-CCM+ Ver.11.04.010-R8. The Eulerian multiphase approach was employed, in combination with a baseline ‘Zero Closure model’ for two-phase flow boiling, applicable to high void fraction regimes. The capability and practicability of MCFD for BWR fuel assembly were evaluated against the international OECD/NRC BWR Full-size Finemesh Bundle Test (BFBT) benchmark data. The benchmark selected, BFBT-4101-61, is a steady-state full power case, with high void fraction conditions. Geometrical modeling and mesh construction of BWR assemblies introduce challenges in generating high quality computational meshes, while retaining a low cell count to allow acceptable runtimes. In order to evaluate the scalability of the CFD code, parallel CFD simulations were carried out with the High Performance Computing (HPC) system of the Idaho National Laboratory. M-CFD results display good agreement with the experimental data for both local void fraction distribution and exit quality. The parallel speedup factor is linearly proportional to the number of processors up to 500 cores, and retains over 92% efficiency with 1000 cores. The execution time of the simulation decreased exponentially with increasing number of processors while the cost of the simulation increased linearly. The proposed baseline closure, in combination with the parallel performance of the STAR-CCM+ code demonstrated promising capabilities for analyzing BWR fuel assemblies.

97 MATHEMATICS AND COMPUTING↗

A Parallel Graph Environment for Real-World Data Analytics Workflows

Economic competitiveness and national security depend increasingly on the insightful analysis of large data sets. The diversity of real-world data sources and analytic workflows impose challenging hardware and software requirements for parallel graph platforms. The irregular nature of graph methods is not supported well by the deep memory hierarchies of conventional distributed systems, requiring new processor and runtime system designs to tolerate memory and synchronization latencies. Moreover, the efficiency of relational table operations and matrix computations are not attainable when data is stored in common graph data structures. In this paper, we present HAGGLE, a high-performance, scalable data analytics platform. The platform’s hybrid data model supports a variety of distributed, thread-safe data structures, parallel programming constructs, and persistent and streaming data. An abstract runtime layer enables us to map the stack to conventional, distributed computer systems with accelerators. The runtime uses multithreading, active messages, and data aggregation to hide memory and synchronization latencies on large-scale systems.

Graph Analytics↗