Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “distributed and parallel processing”

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 523 records · Page 29
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Scalable balanced training of conditional generative adversarial neural networks on image data

Here, we propose a distributed approach to train deep convolutional generative adversarial neural network (DC-CGANs) models. Our method reduces the imbalance between generator and discriminator by partitioning the training data according to data labels, and enhances scalability by performing a parallel training where multiple generators are concurrently trained, each one of them focusing on a single data label. Performance is assessed in terms of inception score, Fréchet inception distance, and image quality on MNIST, CIFAR10, CIFAR100, and ImageNet1k datasets, showing a significant improvement in comparison to state-of-the-art techniques to training DC-CGANs. Weak scaling is attained on all the four datasets using up to 1000 processes and 2000 NVIDIA V100 GPUs on the OLCF supercomputer Summit.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Optimizing the Accelerated Recursive Doubling Algorithm for Block Tridiagonal Systems of Equations

The need to solve block tridiagonal systems with hundreds or thousands of right-hand sides for the same block tridiagonal matrix is common in a variety of disciplines. To meet this need, the Accelerated Recursive Doubling Algorithm was developed. After a right-hand side independent phase, the algorithm allows for the quick, online calculation of solutions for different right-hand sides. In this work, we present methods to optimize the Accelerated Recursive Doubling Algorithm in memory usage and computation time in a hybrid parallelization model. The right-hand side independent phase of the naïve implementation takes ≥ 11/3 the amount of memory required to store the tridiagonal matrix, while our implementation reduces the fraction to ≈ 5/3 . The right-hand side dependent phase of the naïve implementation takes ≥ 6 times the amount of memory required to store the right-hand side, while our implementation reduces the fraction to ≈ 3. The computation time for the independent phase is reduced to ≈ 2/3 times that of the naïve implementation, while the computation time for the dependent phase is reduced to ≈ 5/9 . With increasing numbers of shared-memory threads q on every distributed processing element, we have O(q) theoretical speedup.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1995

Visualization and Tracking of Parallel CFD Simulations

We describe a system for interactive visualization and tracking of a 3-D unsteady computational fluid dynamics (CFD) simulation on a parallel computer. CM/AVS, a distributed, parallel implementation of a visualization environment (AVS) runs on the CM-5 parallel supercomputer. A CFD solver is run as a CM/AVS module on the CM-5. Data communication between the solver, other parallel visualization modules, and a graphics workstation, which is running AVS, are handled by CM/AVS. Partitioning of the visualization task, between CM-5 and the workstation, can be done interactively in the visual programming environment provided by AVS. Flow solver parameters can also be altered by programmable interactive widgets. This system partially removes the requirement of storing large solution files at frequent time steps, a characteristic of the traditional 'simulate (yields) store (yields) visualize' post-processing approach.

Vaziri, Arsi↗
DOE OSTI2022

Hvac: Removing I/O Bottleneck for Large-Scale Deep Learning Applications

Scientific communities are increasingly adopting deep learning (DL) models in their applications to accelerate scientific discovery processes. However, with rapid growth in the computing capabilities of HPC supercomputers, large-scale DL applications have to spend a significant portion of training time performing I/O to a parallel storage system. Previous research works have investigated optimization techniques such as prefetching and caching. Unfortunately, there exist non-trivial challenges to adopting the existing solutions on HPC supercomputers for large-scale DL training applications, which include non-performance and/or failures at extreme scale, lack of portability and generality in design, complex deployment methodology, and being limited to a specific application or dataset. To address these challenges, we propose High-Velocity AI Cache (HVAC), a distributed read-cache layer that targets and fully exploits the node-local storage or near node-local storage technology. HVAC seamlessly accelerates read I/O by aggregating node-local or near node-local storage, avoiding metadata lookups and file locking while preserving portability in the application code. We deploy and evaluate HVAC on 1,024 nodes (with over 6000 NVIDIA V100 GPUS) of the Summit supercomputer. In particular, we evaluate the scalability, efficiency, accuracy, and load distribution of HVAC compared to GPFS and XFS-on-NVMe. With four different DL applications, we observe an average 25 % performance improvement atop GPFS and 9% drop against XFS-on-NVMe, which scale linearly and are considered the performance upper bound. We envision HVAC as an important caching library for upcoming HPC supercomputers such as Frontier.

Khan, Awais↗
NASA NTRS1975

Study of growth of single crystal ribbon in space

The technical feasibility is studied of growing single-crystal silicon ribbon in the space environment. Procedures are described for calculating the electromagnetic fields produced in a silicon ribbon by an rf shaping coil. The forces on the ribbon and the degree of shaping to be expected are determined. The expected steady-state temperature distribution in the ribbon is calculated in the one-dimensional approximation. Calculations on simplified models indicate, that lack of flatness of the shaped ribbon and excessive heating of the melt by the eddy currents induced by the shaping fields may pose problems. An analysis of the relative effects of various kinds of forces other than electromagnetic showed that in the space environment capillarity forces would dominate, and that the shape of the melt is thus principally determined by the shape of any solids with which it comes in contact. This suggests that ribbon may be produced simply by drawing between parallel wires. A concept is developed for a process of off-angle growth, in which the ribbon is pulled at an angle to the solidification front. Such a process promises to offer increased growth rate, better homogeneity, and thinner ribbon.

Wood, V. E.↗
NASA NTRS1999

Data Structure and Parallel Decomposition Considerations on a Fibonacci Grid

The Fibonacci grid, proposed by Swinbank and Purser (see companion abstract), provides attractive properties for global numerical atmospheric prediction by offering an optimally homogeneous, geometrically regular, and approximately isotropic discretization, with only the polar regions requiring special numerical treatment. It is a mathematical idealization, applied to the sphere, of the multi-spiral patterns often found in botanical structures, such as in pine cones and sunflower heads. Computationally, it is natural to organize the domain, into zones, in each of which the same pair, or triple, of "Fibonacci spirals" dominate. But the further subdivision of such zones into "tiles" of a shape and size suitable for distribution to the processors of a massively parallel computer requires very careful consideration if the subsequent spatial computations along the respective spirals, especially those computations (such as compact differencing schemes) that involve recursion, can be implemented in an efficient "load-balanced "manner without requiring excessive amounts of inter-processor communications. In this paper we show how certain "number theoretic" properties of the Fibonacci sequence (whose numbers prescribe the multiplicity of successive spirals) may be exploited in the decomposition of grid zones into tidy arrangements of triangular grid tiles, each tile possessing one side approximately parallel to the constant-latitude zone boundary. We also describe how the spatially recursive processes may be decomposed across such a tiling, and the directionality of the recursions reversed on alternate grid lines, to ensure a very high degree of load balancing throughout the execution of the computations required for one time step of a global model.

Michalakes, John↗
NASA NTRS2010

Collective Intelligence and Three Aspects of Planning in Organizations: A NASA Example

For many complex sociotechnical systems, planning seems to require explicit coordination; certainly, in executing a plan the activities of different actors must be tightly coupled. However, distributing the needed planning information can be very burdensome and error prone, because different groups need different collections of information, updated or kept current on different time cycles. Further, the information needed to form successful plans is often highly distributed, and while feedback about the success of prior plans may exist, it may not be available to those in a position of using this to improve plans or to detect and resolve other problems in the system (Weick, 1995). Tools to support various aspects of planning have been developed, and can provide a huge benefit to the individuals working on that aspect. To be tractable, most solutions address a quite bounded slice of work, isolating it from the larger context. Prospective planning takes place over multiple, nested cycles of decision making. This builds a plan that specifies activities of different granularity. "Subplans" may specify multiple parallel activities by different groups and individuals, as well as sequential, nested actions by a single actor. Planning produces valuable, sharable, external, representations: in addition to prospective use, plans support retrospective assessment and also action in the present. Viewing planning in a larger context - both temporal and organizational -- enables noticing what one does not know and generating more systemic and effective solutions. Viewing a problem as one of collective intelligence invites thinking about the larger organizational context. Many approaches to supporting collective intelligence do not support execution of highly contingent actions, distributed across many players, and hence provide incomplete support for planning. However, CI technology maybe helpful in managing the processes of gathering information for decision making in planning and of distributing plan information to various actors, needed on various time cycles.

Billman, Dorrit↗
NASA NTRS2013

Modeling of Particle Acceleration at Multiple Shocks via Diffusive Shock Acceleration: Preliminary Results

Successful forecasting of energetic particle events in space weather models require algorithms for correctly predicting the spectrum of ions accelerated from a background population of charged particles. We present preliminary results from a model that diffusively accelerates particles at multiple shocks. Our basic approach is related to box models in which a distribution of particles is diffusively accelerated inside the box while simultaneously experiencing decompression through adiabatic expansion and losses from the convection and diffusion of particles outside the box. We adiabatically decompress the accelerated particle distribution between each shock by either the method explored in Melrose and Pope (1993) and Pope and Melrose (1994) or by the approach set forth in Zank et al. (2000) where we solve the transport equation by a method analogous to operator splitting. The second method incorporates the additional loss terms of convection and diffusion and allows for the use of a variable time between shocks. We use a maximum injection energy (E(sub max)) appropriate for quasi-parallel and quasi-perpendicular shocks and provide a preliminary application of the diffusive acceleration of particles by multiple shocks with frequencies appropriate for solar maximum (i.e., a non-Markovian process).

Parker, L. Neergaard↗
NASA NTRS2026

Influence of Processing Parameters on the Mechanical Properties of 3D Printed Borosilicate Particulate Reinforced Polymer Composites

Emerging composite materials are expanding the potential of additive manufacturing and enabling applications previously restricted by traditional manufacturing methods. The multi-phase nature of these composite materials combined with the complex in-ternal geometry of additively manufactured parts have enabled unique behavior, and potentially new applications. Additionally, these materials can be pyrolyzed to create dense metal, ceramic, and glass parts with geometries typically not achievable by tra-ditional processes. Additive manufacturing of borosilicate glass-based systems can open new applications in nuclear engineering, astronomy, and bone regrowth therapy. To elucidate the process-parameter relationship of borosilicate-polylactic acid (PLA) composites, mechanical testing was conducted and compared with a pure polylactic acid polymer baseline. Test specimens were fabricated by fused-filament fabrication with minimal post-processing. Yield strength, ultimate strength, and elastic modulus were calculated from stress-strain curves. Optical and scanning electron microscopy were conducted to observe the specimen microstructure before and after testing. The highest compressive yield strength for the composite was 28.22 MPa, and the highest compressive yield strength for PLA was 49.30 MPa. Print orientation was found to benefit the composite material but have a detrimental effect on the pure matrix material. An elastic modulus of 2.66 GPa was recorded for the borosilicate-PLA composite at 100% infill, 1 shell wall, and layer lines parallel to compression axis. Microscopy revealed that lower modulus composite specimens had the particulates re-distributed within the matrix. Tensile testing was done according to a polymer testing standard, which caused difficulties obtaining consistent fracture within the gauge length.

mechanical testing↗
DOE OSTI2024

Boundary of the Distribution of Solar Wind Proton Beta versus Temperature Anisotropy

Abstract The frequency distribution of solar wind protons, measured in the vicinity of Earth’s orbit, is customarily plotted in ( β ∥ , T ⊥ / T ∥ ) phase space. Here, T ⊥ / T ∥ is the ratio of perpendicular and parallel temperatures, and β ∥ = 8 π nT ∥ / B 2 is the ratio of parallel thermal energy to background magnetic field energy, the so-called “parallel beta,” with ⊥ and ∥ denoting directions with respect to the ambient magnetic field. Such a frequency distribution, plotted as a two-dimensional histogram, forms a peculiar rhombic shape defined with an outer boundary in the said phase space. Past studies reveal that the threshold conditions for temperature anisotropy–driven plasma instability partially account for the boundary on the high- β ∥ side. The low- β ∥ side remains largely unexplained despite some efforts. Work by Vafin et al. recently showed that certain contours of collisional relaxation frequency, ν pp , when parameterized by T ⊥ / T ∥ and β ∥ , could match the overall shape of the left-hand boundary, thus suggesting that the collisional relaxation process might be closely related to the formation of the left-hand boundary. The present paper extends the analysis by Vafin et al. and carries out the dynamical computation of the collisional relaxation process for an ensemble of initial proton states with varying degrees of anisotropic temperatures. The final states of the relaxed protons are shown to closely match the observed boundary to the left of the ( β ∥ , T ⊥ / T ∥ ) phase space. When coupled with a similar set of calculations for the ensemble in the collective instability regime, it is found that the combined collisional/collective effects provide the baseline explanation for the observation.

Astronomy & Astrophysics↗
NASA NTRS1999

Orientation Effects in ZnO Films Using Off-Axis Sputtering Deposition

ZnO is a wide-band-gap oxide material and has been used in numerous applications. It is also a good substrate for fabricating GaN-based (a blue laser candidate) devices. Off-axis sputtering technique is one of the best techniques in synthesizing oxide materials because negative ion bombardment and particle kinetic energy is greatly reduced when adatoms condense on substrates. Since the sputtered material from the target arrive on the substrate surface at a 90 deg. configuration, which differs from the normal sputtering geometry, it is expected that the film uniformity and composition distributions will be affected. However, the details of these properties and mechanisms have not been well studied. ZnO films are synthesized on (0001) sapphire and quartz substrates by off-axis sputtering deposition in various oxygen/argon mixture ratios and pressures at different temperatures. Substrates and sputtering sources are placed at three different orientations that are orthogonal to each other. The normal direction of a substrate is parallel to the gravity vector and the other is perpendicular to it. Film thickness profiles at different growth orientations are determined using a profimeter. All films grown at high temperatures have highly textured structures on quartz substrates and epitaxially grow on sapphire substrates. Because of this process, the film surface is very smooth. X-ray diffraction, scanning probe microscopy, and Fourier transfer infrared spectroscopy, and electrical measurements will be used to characterize these films. Detailed results will be discussed in the presentation.

Zhu, Shen↗
NASA NTRS2009

Model Calibration in Watershed Hydrology

Hydrologic models use relatively simple mathematical equations to conceptualize and aggregate the complex, spatially distributed, and highly interrelated water, energy, and vegetation processes in a watershed. A consequence of process aggregation is that the model parameters often do not represent directly measurable entities and must, therefore, be estimated using measurements of the system inputs and outputs. During this process, known as model calibration, the parameters are adjusted so that the behavior of the model approximates, as closely and consistently as possible, the observed response of the hydrologic system over some historical period of time. This Chapter reviews the current state-of-the-art of model calibration in watershed hydrology with special emphasis on our own contributions in the last few decades. We discuss the historical background that has led to current perspectives, and review different approaches for manual and automatic single- and multi-objective parameter estimation. In particular, we highlight the recent developments in the calibration of distributed hydrologic models using parameter dimensionality reduction sampling, parameter regularization and parallel computing.

Yilmaz, Koray K.↗
NASA NTRS2010

The OMPS Limb Profiler Instrument: Two-Dimensional Retrieval Algorithm

The upcoming Ozone Mapper and Profiler Suite (OMPS), which will be launched on the NPOESS Preparatory Project (NPP) platform in early 2011, will continue monitoring the global distribution of the Earth's middle atmosphere ozone and aerosol. OMPS is composed of three instruments, namely the Total Column Mapper (heritage: TOMS, OMI), the Nadir Profiler (heritage: SBUV) and the Limb Profiler (heritage: SOLSE/LORE, OSIRIS, SCIAMACHY, SAGE III). The ultimate goal of the mission is to better understand and quantify the rate of stratospheric ozone recovery. The focus of the paper will be on the Limb Profiler (LP) instrument. The LP instrument will measure the Earth's limb radiance (which is due to the scattering of solar photons by air molecules, aerosol and Earth surface) in the ultra-violet (UV), visible and near infrared, from 285 to 1000 nm. The LP simultaneously images the whole vertical extent of the Earth's limb through three vertical slits, each covering a vertical tangent height range of 100 km and each horizontally spaced by 250 km in the cross-track direction. Measurements are made every 19 seconds along the orbit track, which corresponds to a distance of about 150km. Several data analysis tools are presently being constructed and tested to retrieve ozone and aerosol vertical distribution from limb radiance measurements. The primary NASA algorithm is based on earlier algorithms developed for the SOLSE/LORE and SAGE III limb scatter missions. All the existing retrieval algorithms rely on a spherical symmetry assumption for the atmosphere structure. While this assumption is reasonable in most of the stratosphere, it is no longer valid in regions of prime scientific interest, such as polar vortex and UTLS regions. The paper will describe a two-dimensional retrieval algorithm whereby the ozone distribution is simultaneously retrieved vertically and horizontally for a whole orbit. The retrieval code relies on (1) a forward 2D Radiative Transfer code (to model limb radiances within a non-uniform atmosphere and evaluate 2D analytical partial derivatives) and (2) an optimal estimator inversion routine. The algorithm uses the typically sparse nature of the kernel matrices as well as fast matrix inversion techniques to allow for fast inversion of limb data with efficient memory management (as was done for MIPAS data processing). While the method has so far only been developed in the context of Single Scatter, the paper will show how the CPU intensive Multiple Scatter modeling can be implemented using parallel CPU processing. Initial results will be presented in terms of retrieved ozone profiles and code performance.

Rault, Didier F.↗
DOE OSTI2026

PRISMA: PARALLEL REFINEMENT AND INTEGRATION SYSTEM FOR MULTI-AZIMUTHAL ANALYSIS

The Parallel Refinement and Integration System for Multi-azimuthal Analysis (PRISMA, version 1.1.0) is a Python application for processing X-ray diffraction (XRD) image data. PRISMA wraps GSAS-II to perform azimuthally-binned peak refinement, computes per-frame strain and d-spacing from those fits, and provides three PyQt5 graphical interfaces: (1) a Recipe Builder for selecting GSAS-II control (.imctrl) files, optional mask (.immask) files or threshold-ased masking, reference and experiment image sets, peaks, zimuthal range and bin size, and an optional ceria-based auto-calibration; (2) a Batch Processor that uses Dask on local workstations and pure MPI (mpi4py.futures.MPICommExecutor) on HPC to distribute GSAS-II refinement across cores or compute nodes and write results to a 4-dimensional (peaks x frames x azimuths x measurements) Zarr dataset; and (3) a Data Analyzer that renders heatmaps of fit parameters, strain, frame-to-frame deltas, and percent-change-vs-reference, and exports user-defined subsections to CSV or Excel. The peak-refinement algorithm is deterministic. Benchmark on ALCF Crux: a 20,000-image set, single-peak fit in frame mode with 44 azimuthal bins on 128 nodes x 128 workers, 48 seconds total wall time.

Lorenzo Martin, Maria De La Cinta [Argonne Nationa↗
NASA NTRS2019

Experimentally-Derived Phase Function Approximations in Support of the Orbital Debris Program Office

The NASA Orbital Debris Program Office (ODPO) has used various optical assets to acquire photometric data of Earth-orbiting objects to define the orbital debris environment. To better characterize and model optical data acquired from ground-based telescopes, the Optical Measurements Center (OMC) at NASA Johnson Space Center emulates illumination conditions seen in space by using equipment and techniques that parallel telescopic observations and source-target-sensor orientations. One of the OMC goals is to improve the size calculation used for optical data by developing an optical-based Size Estimation Model. The current size estimation requires applying a Lambertian phase function, a set albedo value, and range to the observed magnitude. The first step to improving the sampled brightness of laboratory targets is to remove aspect-angle dependencies. Then, the volume of possible object viewing angles is sampled at 21 combinations of azimuth and elevation angles for each solar phase angle. Finally, the acquired images are input into an image processing program that generates approximations for the object’s Bidirectional Reflectance Distribution Function (BRDF) and phase function. The BRDF is a radiometric concept that identifies an object’s material composition by matching a BRDF approximated with photometric data collected by ground-based telescopes with a BRDF generated experimentally from a known object in the laboratory. This paper presents the initial BRDF and phase function approximations for various fragments/targets acquired in the OMC and how the findings will be incorporated into ODPO models. A Lambertian sphere is used as a baseline for initial size estimation calculations and phase function comparisons. Spacecraft materials and fragments from hypervelocity laboratory impact tests are also presented to compare against the current assumed Lambertian phase function used for size estimates. This paper presents the preliminary phase function analysis and plan forward to utilize a laboratory-based phase function to improve the current optical size estimates using BRDF measurements for a large volume of targets composed of various shapes, sizes, and materials.

Hostetler, J.↗
NASA NTRS1979

Charged dust in the outer planetary magnetospheres. I - Physical and dynamical processes

Typical interplanetary grains (Brownlee types) and the physical and dynamical processes associated with their entry into the Jovian magnetosphere are discussed. The charging and subsequent distribution of interplanetary dust grains entering the magnetosphere as well as the basic aspects of orbital dynamics of charged grains are examined. Results showed that all negatively charged fragments are strongly attracted towards the planet by the radial corotational electric field, while some are stably trapped, suggesting that the sudden enhancement by about two orders of magnitude of the interplanetary dust flux, measured at 30 Jupiter radii by Pioneer 10, is a combination of these two results. The similarity in brightness asymmetries between the Jovian and Saturnian satellites leads to the expectation that Saturn's magnetic moment and spin are parallel with the limit of plasma corotational lying between satellites Rhea and Iapetus.

Hill, J. R.↗
NASA NTRS1987

Particle simulation of plasmas on the massively parallel processor

Particle simulations, in which collective phenomena in plasmas are studied by following the self consistent motions of many discrete particles, involve several highly repetitive sets of calculations that are readily adaptable to SIMD parallel processing. A fully electromagnetic, relativistic plasma simulation for the massively parallel processor is described. The particle motions are followed in 2 1/2 dimensions on a 128 x 128 grid, with periodic boundary conditions. The two dimensional simulation space is mapped directly onto the processor network; a Fast Fourier Transform is used to solve the field equations. Particle data are stored according to an Eulerian scheme, i.e., the information associated with each particle is moved from one local memory to another as the particle moves across the spatial grid. The method is applied to the study of the nonlinear development of the whistler instability in a magnetospheric plasma model, with an anisotropic electron temperature. The wave distribution function is included as a new diagnostic to allow simulation results to be compared with satellite observations.

Gledhill, I. M. A.↗
DOE OSTI2020

Scalable Knowledge Graph Analytics at 136 Petaflop/s

We are motivated by newly proposed methods for data mining large-scale corpora of scholarly publications, such as the full biomedical literature, which may consist of tens of millions of papers spanning decades of research. In this setting, analysts seek to discover how concepts relate to one another. They construct graph representations from annotated text databases and then formulate the relationship-mining problem as one of computing all-pairs shortest paths (APSP), which becomes a significant bottleneck. In this context, we present a new high-performance algorithm and implementation of the Floyd-Warshall algorithm for distributed-memory parallel computers accelerated by GPUs, which we call DSNAPSHOT (Distributed Accelerated Semiring All-Pairs Shortest Path). For our largest experiments, we ran DSNAPSHOT on a connected input graph with millions of vertices using 4, 096nodes (24,576GPUs) of the Oak Ridge National Laboratory's Summit supercomputer system. We find DSNAPSHOT achieves a sustained performance of 136×1015 floating-point operations per second (136petaflop/s) at a parallel efficiency of 90% under weak scaling and, in absolute speed, 70% of the best possible performance given our computation (in the single-precision tropical semiring or “min-plus” algebra). Looking forward, we believe this novel capability will enable the mining of scholarly knowledge corpora when embedded and integrated into artificial intelligence-driven natural language processing workflows at scale.

Kannan, Ramakrishnan {ramki}↗