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 775 records · Page 43

Computing contingency statistics in parallel : design trade-offs and limiting cases.

Statistical analysis is typically used to reduce the dimensionality of and infer meaning from data. A key challenge of any statistical analysis package aimed at large-scale, distributed data is to address the orthogonal issues of parallel scalability and numerical stability. Many statistical techniques, e.g., descriptive statistics or principal component analysis, are based on moments and co-moments and, using robust online update formulas, can be computed in an embarrassingly parallel manner, amenable to a map-reduce style implementation. In this paper we focus on contingency tables, through which numerous derived statistics such as joint and marginal probability, point-wise mutual information, information entropy, and {chi}{sup 2} independence statistics can be directly obtained. However, contingency tables can become large as data size increases, requiring a correspondingly large amount of communication between processors. This potential increase in communication prevents optimal parallel speedup and is the main difference with moment-based statistics (which we discussed in [1]) where the amount of inter-processor communication is independent of data size. Here we present the design trade-offs which we made to implement the computation of contingency tables in parallel. We also study the parallel speedup and scalability properties of our open source implementation. In particular, we observe optimal speed-up and scalability when the contingency statistics are used in their appropriate context, namely, when the data input is not quasi-diffuse.

99 GENERAL AND MISCELLANEOUS↗

PARTICLE ACCELERATION AT QUASI-PARALLEL SHOCK WAVES: THEORY AND OBSERVATIONS AT 1 AU

In this paper, we describe a theoretical model for accelerating an arbitrary upstream particle distribution. Only those particles that exceed a prescribed injection energy, E{sub inj}, are accelerated via the diffusive shock acceleration (DSA) mechanism, also known as first-order Fermi acceleration. We identify a set of quasi-parallel shocks at 1 AU and use the observed solar wind particle distribution information to construct our upstream distribution, which is then accelerated diffusively at the shock, assuming the observed shock parameters. The injection energy for particles to be accelerated diffusively at a quasi-parallel shock is discussed theoretically. By using the observed upstream solar wind distribution function and the observed shock parameters, we can compute the injection energy that matches the observed downstream accelerated particle spectrum. Like the previous studies of van Nes et al., Lario et al., and Ho et al., this analysis focuses on the acceleration of protons only via the first-order Fermi acceleration mechanism. However, our primary focus is on quasi-parallel shocks and the injection mechanism in the context of DSA with a background thermal solar wind modeled as a Maxwellian or kappa distribution. Our approach allows for a direct test of injection at interplanetary shocks. It has been proposed that an additional seed population of energetic particles is needed to explain the accelerated particle distribution downstream of quasi-parallel shocks. This conclusion is based typically on studies that address the acceleration of heavy ions primarily and do not characterize the injection of protons alone using the DSA mechanism. Through comparisons of Maxwellian and kappa upstream distributions, we find that DSA with injection directly from a thermal Maxwellian distribution, or weak departures therefrom, for protons is responsible for energetic solar particle events associated with quasi-parallel shocks.

79 ASTRONOMY AND ASTROPHYSICS↗

Modeling the spectral modification of lower hybrid wave in the presence of drift-wave type density fluctuation in the scrape-off-layer of the EAST tokamak

In this work, the spectrum change of the lower hybrid (LH) waves caused by a low-frequency density fluctuation in the scrape-off-layer is studied by applying the wave scattering model developed by Bonoli and Ott [Phys. Fluids 25, 359 (1982)] via a Monte Carlo method. A ray-tracing model specific to this purpose is developed to evaluate the probability distribution of both the poloidal refractive index (N θ ) and the parallel refractive index (N ∥ ) of the LH wave at the last closed flux surface (LCFS).The probability distributions of N ∥ and N θ at the LCFS are studied using the EAST parameters as a function of wave frequency, the initial N ∥ , and the polar injection position, which may influence the lower hybrid current drive efficiency.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

SU-F-SPS-05: Experimental Validation of Peripheral Dose Distribution of Electron Beams for Eclipse Electron Monte Carlo Algorithm

Purpose: In this study, the two available calculation algorithms of the Varian Eclipse treatment planning system(TPS), the electron Monte Carlo(eMC) and General Gaussian Pencil Beam(GGPB) algorithms were used to compare measured and calculated peripheral dose distribution of electron beams. Methods: Peripheral dose measurements were carried out for 6, 9, 12, 15, 18 and 22 MeV electron beams of Varian Triology machine using parallel plate ionization chamber and EBT3 films in the slab phantom. Measurements were performed for 6×6, 10×10 and 25×25cm{sup 2} cone sizes at dmax of each energy up to 20cm beyond the field edges. Using the same film batch, the net OD to dose calibration curve was obtained for each energy. Films were scanned 48 hours after irradiation using an Epson 1000XL flatbed scanner. Dose distribution measured using parallel plate ionization chamber and EBT3 film and calculated by eMC and GGPB algorithms were compared. The measured and calculated data were then compared to find which algorithm calculates peripheral dose distribution more accurately. Results: The agreement between measurement and eMC was better than GGPB. The TPS underestimated the out of field doses. The difference between measured and calculated doses increase with the cone size. The largest deviation between calculated and parallel plate ionization chamber measured dose is less than 4.93% for eMC, but it can increase up to 7.51% for GGPB. For film measurement, the minimum gamma analysis passing rates between measured and calculated dose distributions were 98.2% and 92.7% for eMC and GGPB respectively for all field sizes and energies. Conclusion: Our results show that the Monte Carlo algorithm for electron planning in Eclipse is more accurate than previous algorithms for peripheral dose distributions. It must be emphasized that the use of GGPB for planning large field treatments with 6 MeV could lead to inaccuracies of clinical significance.

60 APPLIED LIFE SCIENCES↗

Scalable Parallel Methods for Analyzing Metagenomics Data at Extreme Scale

The field of bioinformatics and computational biology is currently experiencing a data revolution. The exciting prospect of making fundamental biological discoveries is fueling the rapid development and deployment of numerous cost-effective, high-throughput next-generation sequencing technologies. The result is that the DNA and protein sequence repositories are being bombarded with new sequence information. Databases are continuing to report a Moore’s law-like growth trajectory in their database sizes, roughly doubling every 18 months. In what seems to be a paradigm-shift, individual projects are now capable of generating billions of raw sequence data that need to be analyzed in the presence of already annotated sequence information. While it is clear that data-driven methods, such as sequencing homology detection, are becoming the mainstay in the field of computational life sciences, the algorithmic advancements essential for implementing complex data analytics at scale have mostly lagged behind. Sequence homology detection is central to a number of bioinformatics applications including genome sequencing and protein family characterization. Given millions of sequences, the goal is to identify all pairs of sequences that are highly similar (or “homologous”) on the basis of alignment criteria. While there are optimal alignment algorithms to compute pairwise homology, their deployment for large-scale is currently not feasible; instead, heuristic methods are used at the expense of quality. In this dissertation, we present the design and evaluation of a parallel implementation for conducting optimal homology detection on distributed memory supercomputers. Our approach uses a combination of techniques from asynchronous load balancing (viz. work stealing, dynamic task counters), data replication, and exact-matching filters to achieve homology detection at scale. Results for a collection of 2.56M sequences show parallel efficiencies of ~75-100% on up to 8K cores, representing a time-to-solution of 33 seconds. We extend this work with a detailed analysis of single-node sequence alignment performance using the latest CPU vector instruction set extensions. Preliminary results reveal that current sequence alignment algorithms are unable to fully utilize widening vector registers.

59 BASIC BIOLOGICAL SCIENCES↗

Path-accelerated stochastic molecular dynamics: Parallel-in-time integration using path integrals

Massively parallel computer architectures create new opportunities for the performance of long-time scale molecular dynamics (MD) simulations. Here, we introduce the path-accelerated molecular dynamics method that takes advantage of distributed computing to reduce the wall-clock time of MD simulation via parallelization with respect to stochastic MD time steps. The marginal distribution for the time evolution of a system is expressed in terms of a path integral, enabling the use of path sampling techniques to numerically integrate MD trajectories. By parallelizing the evaluation of the path action with respect to time and by initializing the path configurations from a nonequilibrium distribution, the algorithm enables significant speedups in terms of the length of MD trajectories that can be integrated in a given amount of wall-clock time. The method is demonstrated for Brownian dynamics, although it is generalizable to other stochastic equations of motion including open systems. We apply the method to two simple systems, a harmonic oscillator and a Lennard-Jones liquid, and we show that in comparison to the conventional Euler integration scheme for Brownian dynamics, the new method can reduce the wall-clock time for integrating trajectories of a given length by more than three orders of magnitude in the former system and more than two in the latter. This new method for parallelizing MD in the dimension of time can be trivially combined with algorithms for parallelizing the MD force evaluation to achieve further speedup.

Rosa-Raíces, Jorge L.↗

Integrating Asynchronous Task Parallelism with OpenSHMEM, In: OpenSHMEM 2016: OpenSHMEM and Related Technologies. Enhancing OpenSHMEM for Hybrid Environments

Partitioned Global Address Space (PGAS) programming models combine shared and distributed memory features, and provide a foundation for high-productivity parallel programming using lightweight one-sided communications. The OpenSHMEM programming interface has recently begun gaining popularity as a lightweight library-based approach for developing PGAS applications, in part through its use of a symmetric heap to realize more efficient implementations of global pointers than in other PGAS systems. However, current approaches to hybrid inter-node and intra-node parallel programming in OpenSHMEM rely on the use of multithreaded programming models (e.g., pthreads, OpenMP) that harness intra-node parallelism but are opaque to the OpenSHMEM runtime. This OpenSHMEM+X approach can encounter performance challenges such as bottlenecks on shared resources, long pause times due to load imbalances, and poor data locality. Furthermore, OpenSHMEM+X requires the expertise of hero-level programmers, compared to the use of just OpenSHMEM. All of these are hard challenges to mitigate with incremental changes. This situation will worsen as computing nodes increase their use of accelerators and heterogeneous memories.

Computer Science↗

PLATO (Parallel Load Assignment Tool): A Parallel Workload Partitioner for Particle-based Methods

A key first step in many large-scale high-performance scientific computing applications is the distribution of the computational domain across all participating processors of a parallel computing platform such as a shared memory machine, cluster or a tightly-coupled supercomputer. The primary goal of such a partitioning step is to ensure that the computational work performed by each process remains load-balanced throughout the execution of the parallel application. In this project, a general-purpose parallel partitioning tool, code-named PLATO, was designed and implemented. PLATO assumes that the workload is a set of points in an n-dimensional metric space. These points are abstractions of computational units that may represent point-like computational units such as celestial objects in astrophysics, atoms in materials sciences and molecules in biology/chemistry, or may represent more abstract computational units such as grid points in finite difference methods, grid elements in finite element methods or n-dimensional points in the feature space of a host of other applications. PLATO supports several key and useful capabilities, namely: (a) multiple load partitioning strategies, (b) support for arbitrary user-defined processor topology, (c) pre-fetching data from neighboring processors based on user-defined parameters, and (d) support for arbitrary number and types of attributes associated with each point in the computational domain. PLATO is implemented in the C programming language using Message Passing Interface (MPI) for inter-processor communications and its parallel performance tested on up to 48 processors of a shared memory machine. This report includes a detailed performance study of PLATO using parallel speedup and iso-efficiency as the two main metrics of parallel performance. In addition, a user manual and DOXYGEN generated code documentation have been included for completeness.

97 MATHEMATICS AND COMPUTING↗

Full-wave simulations of ICRF heating regimes in toroidal plasmas with non-Maxwellian distribution functions

At the power levels required for significant heating and current drive in magnetically-confined toroidal plasma, modification of the particle distribution function from a Maxwellian shape is likely [T. H. Stix, Nucl. Fusion, 15 737 (1975)], with consequent changes in wave propagation and in the location and amount of absorption. In order to study these effects computationally, both the finite-Larmor-radius and the high-harmonic fast wave (HHFW), versions of the full-wave, hot-plasma toroidal simulation code TORIC [M. Brambilla, Plasma Phys. Control. Fusion 41, 1 (1999) and M. Brambilla, Plasma Phys. Control. Fusion 44, 2423 (2002)], have been extended to allow the prescription of arbitrary velocity distributions of the form f(v||, v_perp, psi , theta). For hydrogen (H) minority heating of a deuterium (D) plasma with anisotropic Maxwellian H distributions, the fractional H absorption varies significantly with changes in parallel temperature but is essentially independent of perpendicular temperature. On the other hand, for HHFW regime with anisotropic Maxwellian fast ion distribution, the fractional beam ion absorption varies mainly with changes in the perpendicular temperature. The evaluation of the wave-field and power absorption, through the full wave solver, with the ion distribution function provided by either aMonte-Carlo particle and Fokker-Planck codes is also examined for Alcator C-Mod and NSTX plasmas. Non-Maxwellian effects generally tends to increase the absorption with respect to the equivalent Maxwellian distribution.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Parallel physics-informed neural networks via domain decomposition

Here we develop a distributed framework for the physics-informed neural networks (PINNs) based on two recent extensions, namely conservative PINNs (cPINNs) and extended PINNs (XPINNs), which employ domain decomposition in space and in time-space, respectively. This domain decomposition endows cPINNs and XPINNs with several advantages over the vanilla PINNs, such as parallelization capacity, large representation capacity, efficient hyperparameter tuning, and is particularly effective for multi-scale and multi-physics problems. Here, we present a parallel algorithm for cPINNs and XPINNs constructed with a hybrid programming model described by MPI + X, where X ∈ {CPUs, GPUs}. The main advantage of cPINN and XPINN over the more classical data and model parallel approaches is the flexibility of optimizing all hyperparameters of each neural network separately in each subdomain. We compare the performance of distributed cPINNs and XPINNs for various forward problems, using both weak and strong scalings. Our results indicate that for space domain decomposition, cPINNs are more efficient in terms of communication cost but XPINNs provide greater flexibility as they can also handle time-domain decomposition for any differential equations, and can deal with any arbitrarily shaped complex subdomains. To this end, we also present an application of the parallel XPINN method for solving an inverse diffusion problem with variable conductivity on the United States map, using ten regions as subdomains.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Anisotropic Heating and Parallel Heat Flux in Electron-only Magnetic Reconnection with Intense Guide Fields

Electron-only reconnection (E-REC) is a process recently observed in the Earth’s magnetosheath, where magnetic reconnection occurs at electron kinetic scales, and ions do not couple to the reconnection process. Electron-only reconnection is likely to have a significant impact on the energy conversion and dissipation of turbulence cascades at kinetic scales in some settings. This paper investigates E-REC under different intensities of strong guide fields (the ratio between the guide field and the in-plane asymptotic field strength is 5, 10 and 20, respectively) via two-dimensional fully kinetic particle-in-cell simulations, focusing on electron heating. The simulations are initialized with a force-free current sheet equilibrium under various intensities of strong guide fields. Similarly to previous experimental studies, electron temperature anisotropy along separatrices is observed, which is found to be mainly caused by the variations of parallel temperature. Both regions of anisotropy and parallel temperature increase/decrease along separatrices become thinner with increasing guide fields. Besides, we find a transition from a quadrupolar to a hexapolar (six-polar) to an octopolar (eight-polar) structure in temperature anisotropy and parallel temperature as the guide field intensifies. Non-Maxwellian electron velocity distribution functions (EVDFs) at different locations in the three simulations are observed. Our results show that parallel electron velocity varies notably with different guide field intensities and finite parallel electron heat flux density is observed. The three simulations exhibit features of the Chew–Goldberger–Low theory, with the level of consistency increasing as the guide field strength increases. This explains the electron parallel temperature variations and the shape of the EVDFs observed along the separatrices. This work may provide insights into the understanding of electron heating and parallel heat flux density in E-REC observed in the turbulent magnetosheath.

79 ASTRONOMY AND ASTROPHYSICS↗

Reciprocal-Space Maps of X-Ray Diffraction Intensity Distribution and Their Relation to the Dislocation Structure of Epitaxial Layers

X-ray diffraction (XRD) in asymmetric Bragg geometry was measured and XRD intensity distribution maps in the reciprocal space were constructed for GaN epitaxial layers with various degrees of structural perfection grown on c-sapphire substrates. It is established that equal-intensity lines (isolines) related to a regular system of perpendicular rectilinear threading dislocations are extended in a direction parallel to the surface. For a more chaotic distribution of dislocations with a large fraction of horizontal fragments, these isolines are rotated toward a direction perpendicular to the reciprocal lattice vector, although they still not attain a limiting position characteristic of the ideal mosaic crystal.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Integrated Model, Batch and Domain Parallelism in Training Neural Networks

We propose a new integrated method of exploiting model, batch and domain parallelism for the training of deep neural networks (DNNs) on large distributed-memory computers using minibatch stochastic gradient descent (SGD). Our goal is to find an efficient parallelization strategy for a fixed batch size using $P$ processes. Our method is inspired by the communication-avoiding algorithms in numerical linear algebra. We see $P$ processes as logically divided into a $P_r \times P_c$ grid where the $P_r$ dimension is implicitly responsible for model/domain parallelism and the $P_c$ dimension is implicitly responsible for batch parallelism. In practice, the integrated matrix-based parallel algorithm encapsulates these types of parallelism automatically. We analyze the communication complexity and analytically demonstrate that the lowest communication costs are often achieved neither with pure model nor with pure data parallelism. We also show how the domain parallel approach can help in extending the theoretical scaling limit of the typical batch parallel method.

97 MATHEMATICS AND COMPUTING↗

Energetic ion, atom, and molecule reactions and excitation in low-current H{sub 2} discharges: H{sub a}lpha Doppler profiles

Absolute spectral emissivities for Doppler broadened H{sub a}lpha profiles are measured and compared with predictions of energetic hydrogen ion, atom, and molecule behavior in low-current electrical discharges in H{sub 2} at very high electric field E to gas density N ratios E/N and low values of Nd, where d is the parallel-plate electrode separation. These observations reflect the energy and angular distributions for the excited atoms and quantitatively test features of multiple-scattering kinetic models in weakly ionized hydrogen in the presence of an electric field that are not tested by the spatial distributions of H{sub a}lpha emission. Absolute spectral intensities agree well with predictions. Asymmetries in Doppler profiles observed parallel to the electric field at 4<=E/N<=20 kTd result primarily from excitation by fast H atoms directed toward the cathode and diffusely reflected from the cathode. (1 Td=10{sup -21} V m{sup 2}) The effects of reflection of hydrogen particles and of changes with cathode material are modeled accurately without adjustable parameters. Maximum measured wavelength shifts result from acceleration of H{sup +} ions and charge transfer to fast H atoms. The Doppler profiles are consistent with models of reactions among H{sup +}, H{sub 2}{sup +}, H{sub 3}{sup +}, H, and H{sub 2} leading to fast H atoms and then fast excited H(n=3) atoms.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

File I/O for MPI Applications in Redundant Execution Scenarios

As multi-petascale and exa-scale high-performance computing (HPC) systems inevitably have to deal with a number of resilience challenges, such as a significant growth in component count and smaller circuit sizes with lower circuit voltages, redundancy may offer an acceptable level of resilience that traditional fault tolerance techniques, such as checkpoint/restart, do not. Although redundancy in HPC is quite controversial due to the associated cost for redundant components, the constantly increasing number of cores-per-processor is tilting this cost calculation toward a system design where computation, such as for redundancy, is much cheaper and communication, needed for checkpoint/restart, is much more expensive. Recent research and development activities in redundancy for Message Passing Interface (MPI) applications focused on availability/reliability models and replication algorithms. This paper takes a first step toward solving an open research problem associated with running a parallel application redundantly, which is file I/O under redundancy. The approach intercepts file I/O calls made by a redundant application to employ coordination protocols that execute file I/O operations in a redundancy-oblivious fashion when accessing a node-local file system, or in a redundancy-aware fashion when accessing a shared networked file system. A proof-of concept prototype is presented and a number of coordination protocols are described and evaluated. The results show the performance impact for redundantly accessing a shared networked file system, but also demonstrate the capability to regain performance by utilizing MPI communication between replicas and parallel file I/O.

99 GENERAL AND MISCELLANEOUS↗

Impact of Spherical Inclusion Mean Chord Length and Radius Distribution on Three-Dimensional Binary Stochastic Medium Particle Transport

We describe a parallel benchmark procedure and numerical results for a three-dimensional binary stochastic medium particle transport benchmark problem. The binary stochastic medium is composed of optically thick spherical inclusions distributed in an optically thin background matrix material. We investigate three sphere mean chord lengths, three distributions for the sphere radii (constant, uniform, and exponential), and six sphere volume fractions ranging from 0.05 to 0.3. For each sampled independent material realization, we solve the associated transport problem using the Mercury Monte Carlo particle transport code. We compare the ensemble-averaged benchmark fiducial tallies of reflection from and transmission through the spatial domain as well as absorption in the spherical inclusion and background matrix materials. For the parameter values investigated, we find a significant dependence of the ensemble-averaged fiducial tallies on both sphere mean chord length and sphere volume fraction, with the most dramatic variation occurring for the transmission through the spatial domain. We find a weaker dependence of most benchmark tally quantities on the distribution describing the sphere radii, provided the sphere mean chord length used is the same in the different distributions. The exponential distribution produces larger differences from the constant distribution than the uniform distribution produces. The transmission through the spatial domain does exhibit a significant variation when an exponential radius distribution is used.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Electron cyclotron current drive modelling with parallel momentum correction for tokamaks and stellarators

The electron cyclotron current drive (ECCD) is described by the solution of a drift-kinetic equation (DKE) with the quasi-linear heating term which is highly localised in phase space. Within the adjoint approach, an equivalent DKE must be solved without this highly localised source, and the ECCD is described by a straightforward convolution in phase space. The parallel momentum correction technique [Maassberg et al., Phys. Plasmas 16, 072504 (2009)] reduces the problem to the solution of a simplified mono-energetic DKE. The mono-energetic distribution functions are calculated by the DKES code [W. I. van Rij and S. P. Hirshman, Phys. Fluids B 1, 563 (1989)]. Parallel momentum correction, requiring only an energy-dependent weighting factor, is applied to these distribution functions allowing for the direct estimation of the EC-driven current. For small yet finite collisionalities, a rather simple model for ECCD is presented. In this approach, the symmetric portion of the electron distribution function with respect to v{sub Parallel-To} (the Ware pinch contribution) is neglected and an 'off-set' only in the passing particle domain adds to the collisionless (anti-symmetric) distribution function. For this approximation, only the mono-energetic parallel conductivity coefficient as function of the collisionality is needed. The impact of small yet finite collisionalities on ECCD is shown for X2- and O2-scenarios at W7-X.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Large-Scale Multi-Dimensional Document Clustering on GPU Clusters

Document clustering plays an important role in data mining systems. Recently, a flocking-based document clustering algorithm has been proposed to solve the problem through simulation resembling the flocking behavior of birds in nature. This method is superior to other clustering algorithms, including k-means, in the sense that the outcome is not sensitive to the initial state. One limitation of this approach is that the algorithmic complexity is inherently quadratic in the number of documents. As a result, execution time becomes a bottleneck with large number of documents. In this paper, we assess the benefits of exploiting the computational power of Beowulf-like clusters equipped with contemporary Graphics Processing Units (GPUs) as a means to significantly reduce the runtime of flocking-based document clustering. Our framework scales up to over one million documents processed simultaneously in a sixteennode GPU cluster. Results are also compared to a four-node cluster with higher-end GPUs. On these clusters, we observe 30X-50X speedups, which demonstrates the potential of GPU clusters to efficiently solve massive data mining problems. Such speedups combined with the scalability potential and accelerator-based parallelization are unique in the domain of document-based data mining, to the best of our knowledge.

99 GENERAL AND MISCELLANEOUS↗