Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel, scaling”

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 1,621 records · Page 90
All sourcesNASA NTRSDOE OSTI
NASA NTRS1995

Analysis of composite ablators using massively parallel computation

In this work, the feasibility of using massively parallel computation to study the response of ablative materials is investigated. Explicit and implicit finite difference methods are used on a massively parallel computer, the Thinking Machines CM-5. The governing equations are a set of nonlinear partial differential equations. The governing equations are developed for three sample problems: (1) transpiration cooling, (2) ablative composite plate, and (3) restrained thermal growth testing. The transpiration cooling problem is solved using a solution scheme based solely on the explicit finite difference method. The results are compared with available analytical steady-state through-thickness temperature and pressure distributions and good agreement between the numerical and analytical solutions is found. It is also found that a solution scheme based on the explicit finite difference method has the following advantages: incorporates complex physics easily, results in a simple algorithm, and is easily parallelizable. However, a solution scheme of this kind needs very small time steps to maintain stability. A solution scheme based on the implicit finite difference method has the advantage that it does not require very small times steps to maintain stability. However, this kind of solution scheme has the disadvantages that complex physics cannot be easily incorporated into the algorithm and that the solution scheme is difficult to parallelize. A hybrid solution scheme is then developed to combine the strengths of the explicit and implicit finite difference methods and minimize their weaknesses. This is achieved by identifying the critical time scale associated with the governing equations and applying the appropriate finite difference method according to this critical time scale. The hybrid solution scheme is then applied to the ablative composite plate and restrained thermal growth problems. The gas storage term is included in the explicit pressure calculation of both problems. Results from ablative composite plate problems are compared with previous numerical results which did not include the gas storage term. It is found that the through-thickness temperature distribution is not affected much by the gas storage term. However, the through-thickness pressure and stress distributions, and the extent of chemical reactions are different from the previous numerical results. Two types of chemical reaction models are used in the restrained thermal growth testing problem: (1) pressure-independent Arrhenius type rate equations and (2) pressure-dependent Arrhenius type rate equations. The numerical results are compared to experimental results and the pressure-dependent model is able to capture the trend better than the pressure-independent one. Finally, a performance study is done on the hybrid algorithm using the ablative composite plate problem. It is found that there is a good speedup of performance on the CM-5. For 32 CPU's, the speedup of performance is 20. The efficiency of the algorithm is found to be a function of the size and execution time of a given problem and the effective parallelization of the algorithm. It also seems that there is an optimum number of CPU's to use for a given problem.

Shia, David↗
NASA NTRS1998

3D Navier-Stokes Time Accurate Solutions Using Multipartitioning Parallel Computation Methodology

A parallel CFD code solving 3D time accurate Navier-Stokes equations with multipartitioning parallel Methodology is being developed in collaboration with Ohio State University within the Air Vehicle Directorate, at Wright Patterson Air Force Base. The advantage of the multipartitioning parallel method is that the domain decomposition will not introduce domain boundaries for the implicit operators. A ring structure data communication is employed so that the implicit time accurate method can be implemented for multi-processors with the same accuracy as for the single processor. No sub-iteration is needed at the domain boundaries. The code has been validated for some typical unsteady flows, which include Coutte Flow, flow passing a cylinder. The code now is being employed for a large scale time accurate wall jet transient flow computation. 'ne preliminary results are promising. The mesh has been refined to capture more details of the flow field. The mesh refinement computation is in progress and would be difficult to successfully implement without the parallel computation techniques used. A modified version of the code with more efficient inversion of the diagonalized block matrix is currently being tested.

Zha, Ge-Cheng↗
DOE OSTI2019

FPGA-Accelerated Machine Learning Inference as a Service for Particle Physics Computing

Large-scale particle physics experiments face challenging demands for high-throughput computing resources both now and in the future. New heterogeneous computing paradigms on dedicated hardware with increased parallelization, such as Field Programmable Gate Arrays (FPGAs), offer exciting solutions with large potential gains. The growing applications of machine learning algorithms in particle physics for simulation, reconstruction, and analysis are naturally deployed on such platforms. We demonstrate that the acceleration of machine learning inference as a web service represents a heterogeneous computing solution for particle physics experiments that potentially requires minimal modification to the current computing model. As examples, we retrain the ResNet-50 convolutional neural network to demonstrate state-of-the-art performance for top quark jet tagging at the LHC and apply a ResNet-50 model with transfer learning for neutrino event classification. Using Project Brainwave by Microsoft to accelerate the ResNet-50 image classification model, we achieve average inference times of 60 (10) ms with our experimental physics software framework using Brainwave as a cloud (edge or on-premises) service, representing an improvement by a factor of approximately 30 (175) in model inference latency over traditional CPU inference in current experimental hardware. A single FPGA service accessed by many CPUs achieves a throughput of 600–700 inferences per second using an image batch of one, comparable to large batch-size GPU throughput and significantly better than small batch-size GPU throughput. Deployed as an edge or cloud service for the particle physics computing model, coprocessor accelerators can have a higher duty cycle and are potentially much more cost-effective.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2019

Polarized Soft X-ray Scattering Reveals Chain Orientation within Nanoscale Polymer Domains

Fully conjugated donor-acceptor block copolymers can serve as the active layer in organic photovoltaics and other organic electronic devices. Additionally, self-assembly into periodic domains and crystallization of the constituent blocks are crucial to enable control of the multiscale structure and consequently electronic properties. Resonant soft X-ray scattering (RSoXS) is an invaluable tool to characterize such materials, where tuning the X-ray energy and polarization reveals molecular orientation and domain spacing. In this work, anisotropic soft X-ray scattering data reveal the type and degree of orientation within conjugated block copolymers composed of poly(3-hexylthiophene-2,5-diyl) and various push-pull copolymers. The maximum anisotropy is observed at the scattering vector corresponding to the spacing between block copolymer domains, which scales with the end-to-end distance of the blocks, as expected. Furthermore, the anisotropy in RSoXS data reveals that the crystalline P3HT blocks orient, on average, parallel or nearly parallel to the block copolymer interface; the average tilt angle between P3HT chains and domain interfaces can be extracted from the dependence of the anisotropy with polar angle. We interpret this angle to correspond to the average tilt of rings in the unit cell, potentially within a chiral mesostructure. Results are corroborated with scattering calculations based on simple model structures.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Taylor limit studies for local helicity injection plasma startup

Solenoid-free tokamak startup techniques can simplify the design and reduce the cost of tokamak-based fusion energy systems. The newly commissioned Pegasus-III spherical tokamak provides a dedicated platform for developing a scalable solenoid-free startup approach. Local helicity injection (LHI) is one promising technique being developed on Pegasus-III, which uses compact edge current sources to drive open field line current I inj that is redistributed via helicity-conserving instabilities to initiate tokamak plasmas. These instabilities relax the system toward a minimum energy state described by Taylor relaxation theory. This relaxation process imposes a global current limit for LHI plasmas, referred to as the Taylor limit I TL ⁠, which can dictate the plasma current I p evolution. Experiments on Pegasus and Pegasus-III show that increasing this limit leads to more effective utilization of the available helicity input and higher I p plasmas. Initial Pegasus-III experiments have verified the expected I TL $\varpropto$ ( I inj B T ) 1/2 scaling holds for B T and I inj up to 0.3 T on axis and 12 kA ⁠, respectively. Additionally, data from different injector arrays show that the width of the injector aperture directly alters I TL and can be leveraged for future injector design optimization. Throughout these scaling studies, discharges were overdriven at I p = I TL by providing excess helicity input. Intermittent bursts of n = 1 magnetic activity consistent with large-scale reconnection events occurred more often as the system was increasingly overdriven leading to a flattening of the λ = μ 0 J ∥ / |B| profile, where J ∥ is the parallel current density.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2019

Combining Tensor Decompositions and Graph Analytics to Provide Cyber Situational Awareness at HPC Scale

This paper describes MADHAT (Multidimensional Anomaly Detection fusing HPC, Analytics, and Tensors), an integrated workflow that demonstrates the applicability of HPC resources to the problem of maintaining cyber situational awareness. MADHAT combines two high-performance packages: ENSIGN for large-scale sparse tensor decompositions and HAGGLE for graph analytics. Tensor decompositions isolate coherent patterns of network behavior in ways that common clustering methods based on distance metrics cannot. Parallelized graph analysis then uses directed queries on a representation that combines the elements of identified patterns with other available information (such as additional log fields, domain knowledge, network topology, whitelists and blacklists, prior feedback, and published alerts) to confirm or reject a threat hypothesis, collect context, and raise alerts. MADHAT was developed using the collaborative HPC Architecture for Cyber Situational Awareness (HACSAW) research environment and evaluated on structured network sensor logs collected from Defense Research and Engineering Network (DREN) sites using HPC resources at the U.S. Army Engineer Research and Development Center DoD Supercomputing Resource Center (ERDC DSRC). To date, MADHAT has analyzed logs with over 650 million entries.

Cyber security, Tensor decomposition, graph analyt↗
DOE OSTI2019

Scalable submicrometer additive manufacturing

High-throughput fabrication techniques for generating arbitrarily complex three-dimensional structures with nanoscale features are desirable across a broad range of applications. Two-photon lithography (TPL)–based submicrometer additive manufacturing is a promising candidate to fill this gap. However, the serial point-by-point writing scheme of TPL is too slow for many applications. Attempts at parallelization either do not have submicrometer resolution or cannot pattern complex structures. We overcome these difficulties by spatially and temporally focusing an ultrafast laser to implement a projection-based layer-by-layer parallelization. This increases the throughput up to three orders of magnitude and expands the geometric design space. We demonstrate this by printing, within single-digit millisecond time scales, nanowires with widths smaller than 175 nanometers over an area one million times larger than the cross-sectional area.

36 MATERIALS SCIENCE↗
DOE OSTI2019

FPGAs as a Service to Accelerate Machine Learning Inference [PowerPoint]

Large-scale particle physics experiments face challenging demands for high-throughput computing resources both now and in the future. New heterogeneous computing paradigms on dedicated hardware with increased parallelization, such as Field Programmable Gate Arrays (FPGAs), offer exciting solutions with large potential gains. The growing applications of machine learning algorithms in particle physics for simulation, reconstruction, and analysis are naturally deployed on such platforms. We demonstrate that the acceleration of machine learning inference as a web service represents a heterogeneous computing solution for particle physics experiments that requires minimal modification to the current computing model. As examples, we retrain the ResNet-50 convolutional neural network to demonstrate state-of-the-art performance for top quark jet tagging at the LHC, and apply a ResNet-50 model with transfer learning for neutrino event classification. Using Project Brainwave b y Microsoft to accelerate the ResNet-50 image classification model, we achieve average inference times of 60 (10) milliseconds with our experimental physics software framework using Brainwave as a cloud (edge or on-premises) service, representing an improvement by a factor of approximately 30 (175) in model inference latency over traditional CPU inference in current experimental hardware. A single FPGA service accessed by many CPUs achieves a throughput of 600--700 inferences per second using an image batch of one, comparable to large batch-size GPU throughput and significantly better than small batch-size GPU throughput. Deployed as an edge or cloud service for the particle physics computing model, coprocessor accelerators can have a higher duty cycle and are potentially much more cost-effective.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
NASA NTRS1972

Study of the techniques feasible for food synthesis aboard a spacecraft

Synthesis of sugars by Ca(OH)2 catalyzed formaldehyde condensation (the formose reaction) has produced branched carbohydrates that do not occur in nature. The kinetics and mechanisms of the homogeneously catalyzed autocatalytic condensation were studied and analogies between homogeneous and heterogeneous rate laws have been found. Aldol condensations proceed simultaneously with Cannizzaro and crossed-Cannizzaro reactions and Lobry de Bruyn-Van Eckenstein rearrangements. The separate steps as well as the interactions of this highly complex reaction system were elucidated. The system exhibits instabilities, competitive catalytic, mass action, and equilibrium phenomena, complexing, and parallel and consecutive reactions. Specific finding that have been made on the problem will be of interest for synthesizing sugars, both for sustained space flight and for large scale food manufacture. A contribution to methodology for studying complex catalyzed reactions and to understanding control of reaction selectivity was a broad goal of the project.

Weiss, A. H.↗
NASA NTRS1982

Tone-excited jet: Theory and experiments

A detailed study to understand the phenomenon of broadband jet-noise amplification produced by upstream discrete-tone sound excitation has been carried out. This has been achieved by simultaneous acquisition of the acoustic, mean velocity, turbulence intensities, and instability-wave pressure data. A 5.08 cm diameter jet has been tested for this purpose under static and also flight-simulation conditions. An open-jet wind tunnel has been used to simulate the flight effects. Limited data on heated jets have also been obtained. To improve the physical understanding of the flow modifications brought about by the upstream discrete-tone excitation, ensemble-averaged schlieren photographs of the jets have also been taken. Parallel to the experimental study, a mathematical model of the processes that lead to broadband-noise amplification by upstream tones has been developed. Excitation of large-scale turbulence by upstream tones is first calculated. A model to predict the changes in small-scale turbulence is then developed. By numerically integrating the resultant set of equations, the enhanced small-scale turbulence distribution in a jet under various excitation conditions is obtained. The resulting changes in small-scale turbulence have been attributed to broadband amplification of jet noise. Excellent agreement has been found between the theory and the experiments. It has also shown that the relative velocity effects are the same for the excited and the unexcited jets.

Ahuja, K. K.↗
NASA NTRS1987

Anomalous transport in discrete arcs and simulation of double layers in a model auroral circuit

The evolution and long-time stability of a double layer (DL) in a discrete auroral arc requires that the parallel current in the arc, which may be considered uniform at the source, be diverted within the arc to charge the flanks of the U-shaped double layer potential structure. A simple model is presented in which this current redistribution is effected by anomalous transport based on electrostatic lower hybrid waves driven by the flank structure itself. This process provides the limiting constraint on the double layer potential. The flank charging may be represented as that of a nonlinear transmission line. A simplified model circuit, in which the transmission line is represented by a nonlinear impedance in parallel with a variable resistor, is incorporated in a one-dimensional simulation model to give the current density at the DL boundaries. Results are presented for the scaling of the DL potential as a function of the width of the arc and the saturation efficiency of the lower hybrid instability mechanism.

Smith, Robert A.↗
NASA NTRS1985

SS focused technology: Gateways and NOS's

The extensions and enhancements of the fiber optic data bus technology supported by the Space Station Focused Technology Program are discussed. This includes the operating software for the network called the Network Operating System (NOS); gateways and bridges for multiple network topologies; and very large scale topology implimentations to shrink the size and power of the Bus Interface Unit (BIU) down to more manageable dimensions. CMOS is being investigated for the lower speed (parallel) logic. Gallium arsendide is being studied for the high speed (serial) logic.

Hartenstein, R.↗
NASA NTRS1987

Dual laser gyro test program

This paper addresses the results of a comparative test evaluation of two similar-sized but otherwise dissimilar ring laser gyroscopes (RLG's). Both units were tested side by side, with the input axes (IA's) parallel, on the same test table. This report describes the rationale and design factors considered important to the test objectives. Emphasis was placed upon the evaluation of the scale factor (SF) linearity, drift rate stability (short term and long term), and derived rate. Conclusions drawn were (1) that dual operation had an observable but insignificant effect on the test results, (2) that the benefits of dual operation outweighed the additional design and analysis efforts required at the outset of the program, and (3) that the performance characteristics of the two test articles were significantly different. The differences between the two test articles (1) led to suggestions which could direct one vendor toward obtaining improved performance and (2) resulted in the creation of a conceptually new method (DELTA TIME-COUNT histogram) for quickly assessing the quality of laser gyros (or any digital sensor device for that matter).

Jones, Malcolm E.↗
NASA NTRS1989

Theoretical and experimental investigation of the aeroelastic stability of an advanced bearingless rotor in hover and forward flight

The aeroelastic stability of a shaft-fixed, 1/8th Froude scaled bearingless rotor was investigated in a series of wind tunnel experiments simulating a wide range of operating conditions. A finite element formulation was used to perform a parallel theoretical analysis, with the goal of determining whether a bearingless rotor system could be made aeroelastically stable without the incorporation of auxilliary dampers. A quick estimate of lag mode damping was provided by a refined moving-block analysis implemented in real time which predicted similar damping values. Model rotor and blade properties were also determined, and these properties were used as inputs for a newly refined bearingless rotor analysis. Predicted results were compared with experimental results in hover and forward flight. Results indicated that soft pitch link stiffness increases pitch-lag coupling and stabilizes lag mode stability in hover and at low advance ratios, but destabilizes at higher advance ratios.

Wang, James M.↗
NASA NTRS1990

Programming in a proposed 9X distributed Ada

The proposed Ada 9X constructs for distribution was studied. The goal was to select suitable test cases to help in the evaluation of the proposed constructs. The examples were to be considered according to the following requirements: real time operation; fault tolerance at several different levels; demonstration of both distributed and massively parallel operation; reflection of realistic NASA programs; illustration of the issues of configuration, compilation, linking, and loading; indications of the consequences of using the proposed revisions for large scale programs; and coverage of the spectrum of communication patterns such as predictable, bursty, small and large messages. The first month was spent identifying possible examples and judging their suitability for the project.

Waldrop, Raymond S.↗
NASA NTRS1991

Energy dissipation in structured electrodynamic environments

It is usually assumed that the profile of the ion Pedersen conductivity determines the altitude dependence of the energy dissipation rate This paper points out the strong altitude dependence of the energy dissipation rate on the spatial scale size of the imposed electric field. To illustrate the importance of such considerations, examples of the ubiquity to electric field structure in the high-latitude ionosphere are shown; this is particularly prominent when the interplanetary magnetic field has a northward component. It is then shown quantitatively how the existence of electric field structure with scale sizes of 10 km or less strongly impacts both the altitude extent over which the electromagnetic energy is dissipated and its partitioning between current systems perpendicular and parallel to the magnetic field.

Heelis, R. A.↗
NASA NTRS1993

A Novel Method for Characterization of Superconductors: Physical Measurements and Modeling of Thin Films

A method for characterization of granular superconducting thin films has been developed which encompasses both the morphological state of the sample and its fabrication process parameters. The broad scope of this technique is due to the synergism between experimental measurements and their interpretation using numerical simulation. Two novel technologies form the substance of this system: the magnetically modulated resistance method for characterizing superconductors; and a powerful new computer peripheral, the Parallel Information Processor card, which provides enhanced computing capability for PC computers. This enhancement allows PC computers to operate at speeds approaching that of supercomputers. This makes atomic scale simulations possible on low cost machines. The present development of this system involves the integration of these two technologies using mesoscale simulations of thin film growth. A future stage of development will incorporate atomic scale modeling.

Kim, B. F.↗
NASA NTRS1994

Shock interactions with magnetized interstellar clouds. 1: Steady shocks hitting nonradiative clouds

We study the interaction of a steady, planar shock with a nonradiative, spherical, interstellar cloud threaded by a uniform magnetic field. For strong shocks, the sonic Mach number scales out, so two parameters determine the evolution: the ratio of cloud to intercloud density, and the Alfven Mach number. We focus on the case with initial field parallel to the shock velocity, though we also present one model with field perpendicular to the velocity. Even with 100 zones per cloud radius, we find that the magnetic field structure converges only at early times. However, we can draw three conclusions from our work. First, our results suggest that the inclusion of a field in equipartition with the preshock medium can prevent the complete destruction of the cloud found in the field-free case recently considered by Klein, McKee, & Colella. Second, the interaction of the shock with the cloud can amplify the magnetic field in some regions up to equipartition with the post-shock thermal pressure. In the parallel-field case, the shock preferentially amplifies the parallel component of the field, creating a 'flux rope,' a linear structure of concentrated magnetic field. The flux rope dominates the volume of amplified field, so that laminar, rather than turbulent, amplification is dominant in this case. Third, the presence of the cloud enhances the production of X-ray and synchrotron emission. The X-ray emission peaks early, during the initial passage of the shock over the cloud, while the synchrotron emission peaks later, when the flow sweeps magnetic field onto the axis between the cloud and the main shock.

Low, Mordecai-Mark Mac↗