Search NASA⌕ Search

SEARCH · Search NASA

Results for “data parallel”

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
DOE OSTI2018

Energy Conversion by Parallel Electric Fields During Guide Field Reconnection in Scaled Laboratory and Space Experiments

Here, we present direct and scaled comparisons between laboratory and in situ space observations of magnetic reconnection with a guide field, comparing results from the Magnetospheric Multiscale Mission and the Magnetic Reconnection eXperiment (MRX). While Magnetospheric Multiscale Mission observations obtain high–resolution and fully kinetic data, MRX observations fully cover the 2–D reconnection plane near the current sheet, removing uncertainties in situating the measurements compared to the reconnection region. Through scaling transformations, we show a quantitative agreement in magnetic field and current density profiles, which agree within a factor of 2 from each other. The introduction of the guide field causes the energy conversion J·E in the current sheet to be dominated by J || E || in both cases. However, parallel electric fields reported by recent spacecraft crossings are significantly (5–10 times) larger than values obtained on MRX, highlighting an important issue for understanding energy conversion by reconnection.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2020

Axial Asymmetry Studies in Gaia Data Release 2 Yield the Pattern Speed of the Galactic Bar

Our recent studies of axial-symmetry breaking in the nearby ($d <3 \,{\rm kpc}$) star counts are sensitive to the distortions of stellar orbits perpendicular and parallel to the orientation of the bar just within and beyond the outer Lindblad resonance (OLR) radius. Using the location of the sign flip in the left-right asymmetry in stars counts about the anticenter line to determine the OLR radius $R_{\rm OLR}$, and treating the bar as if it were a weakly non-axisymmetric effect, we use $R_{\rm OLR}$ and recent measurements of the Galactic rotation curve and the Sun-Galactic-center distance $R_{0}$ to determine the pattern speed $\Omega_{\rm p}$ of the Galactic bar, as well as the Galactic corotation radius $R_{\rm CR}$. After removing the effect of the Large and Small Magellanic clouds from our asymmetry measurement, we find that $R_{\rm OLR}=(0.96 \pm 0.03)R_0 = 7.85 \pm 0.25 \ \rm kpc$, $\Omega_{\rm p} = 49.3 \pm 2.2 \ \rm km \ s^{-1} \ kpc^{-1}$, $R_{\rm CR}=(0.58 \pm 0.04)R_0 = 4.76 \pm 0.27 \ \rm kpc$, revealing, as we shall show, that the Milky Way's bar is likely both weak and fast, though we also note possible evidence for non-steady-state effects in the bar region.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2015

RE-VISIT OF HST FUV OBSERVATIONS OF THE HOT-JUPITER SYSTEM HD 209458: NO Si iii DETECTION AND THE NEED FOR COS TRANSIT OBSERVATIONS

The discovery of O i atoms and C ii ions in the upper atmosphere of HD 209458b, made with the Hubble Space Telescope Imaging Spectrograph (STIS) using the G140L grating, showed that these heavy species fill an area comparable to the planet’s Roche lobe. The derived ∼10% transit absorption depths require super-thermal processes and/or supersolar abundances. From subsequent Cosmic Origins Spectrograph (COS) observations, C ii absorption was reported with tentative velocity signatures, and absorption by Si iii ions was also claimed in disagreement with a negative STIS G140L detection. Here, we revisit the COS data set showing a severe limitation in the published results from having contrasted the in-transit spectrum against a stellar spectrum averaged from separate observations, at planetary phases 0.27, 0.72, and 0.49. We find variable stellar Si iii and C ii emissions that were significantly depressed not only during transit but also at phase 0.27 compared to phases 0.72 and 0.49. Their respective off-transit 7.5% and 3.1% flux variations are large compared to their reported 8.2 ± 1.4% and 7.8 ± 1.3% transit absorptions. Significant variations also appear in the stellar line shapes, questioning reported velocity signatures. We furthermore present archive STIS G140M transit data consistent with no Si iii absorption, with a negative result of 1.7 ± 18.7 including ∼15% variability. Silicon may still be present at lower ionization states, in parallel with the recent detection of extended magnesium, as Mg i atoms. In this frame, the firm detection of O i and C ii implying solar or supersolar abundances contradicts the recent inference of potential 20–125× subsolar metallicity for HD 209458b.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

LATTE: open-source, high-performance traveltime computation, tomography and source location in acoustic and elastic media

Traveltime-based tomography and source location are fundamental approaches for imaging subsurface structures and understanding the spatiotemporal distribution of seismicity from local to global scales. We present an open-source, high-performance framework integrating eikonal equation solvers and adjoint-state theory for traveltime computation, velocity tomography, source location and joint tomography-location in 2-D/3-D acoustic and elastic media. We introduce novel regularization schemes based on total generalized p-variation, structural similarity and multitask machine learning to enhance the fidelity and interpretability of inverted models and source locations. Key features of our implementation also include the ability to leverage both absolute-difference and double-difference traveltime misfits for high-fidelity velocity tomography and source parameter estimation; support for traveltime computation and inversion in diverse 2-D/3-D scenarios with arbitrary source and receiver distributions; and a perturbation-based optimal step-size estimation method to reduce computational costs. In addition, our implementation employs shared-memory and distributed-memory parallelization to provide an efficient solution for traveltime computation, tomography, and source location. In conclusion, we validate the efficacy and accuracy of our approach through multiple synthetic data examples.

58 GEOSCIENCES↗
DOE OSTI2021

Revisiting Huffman Coding: Toward Extreme Performance on Modern GPU Architectures

Today's high-performance computing (HPC) applications are producing vast volumes of data, which are challenging to store and transfer efficiently during the execution, such that data compression is becoming a critical technique to mitigate the storage burden and data movement cost. Huffman coding is arguably the most efficient Entropy coding algorithm in information theory, such that it could be found as a fundamental step in many modern compression algorithms such as DEFLATE. On the other hand, today's HPC applications are more and more relying on the accelerators such as GPU on supercomputers, while Huffman encoding suffers from low throughput on GPUs, resulting in a significant bottleneck in the entire data processing. In this paper, we propose and implement an efficient Huffman encoding approach based on modern GPU architectures, which addresses two key challenges: (1) how to parallelize the entire Huffman encoding algorithm, including codebook construction, and (2) how to fully utilize the high memory-bandwidth feature of modern GPU architectures. The detailed contribution is fourfold. (1) We develop an efficient parallel codebook construction on GPUs that scales effectively with the number of input symbols. (2) We propose a novel reduction based encoding scheme that can efficiently merge the codewords on GPUs. (3) We optimize the overall GPU performance by leveraging the state-of-the-art CUDA APIs such as Cooperative Groups. (4) We evaluate our Huffman encoder thoroughly using six real-world application datasets on two advanced GPUs and compare with our implemented multi-threaded Huffman encoder. Experiments show that our solution can improve the encoding throughput by up to 5.0x and 6.8x on NVIDIA RTX 5000 and V100, respectively, over the state-of-the-art GPU Huffman encoder, and by up to 3.3x over the multi-thread encoder on two 28-core Xeon Platinum 8280 CPUs.

Tian, Jiannan↗
DOE OSTI2021

Investigating the magnetic and atomic interface configuration for a model Fe/CrN bilayer system

A bilayer of iron on chromium nitride (Fe/CrN) is an interesting system for exchange biasing and sensing applications as the Néel temperature of CrN is 280 K and the Curie temperature of Fe is 1043 K. In this paper, we study the crystal and magnetic structures of the Fe/CrN interface at the atomic level. High quality epitaxial Fe/CrN bilayers prepared by molecular beam epitaxy grow in 001 orientation on MgO(001) substrates with uniform layer thicknesses and sharp interfaces. Our data reveal the epitaxial correlation between Fe and CrN crystals and their magnetic structures at the interface. The magnetic anisotropy directions of Fe and CrN are found parallel to [110] MgO . We studied the electronic and magnetic properties of the interface by performing the first-principles total-energy calculations. Here, we present a model that combines the crystal and magnetic structures of the Fe/CrN bilayer and fully explains all results.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2019

Temperature Dependence of the Critical Field of the Organic Superconductor {kappa}-(BEDT-TTF){sub 2}Cu[N(CN){sub 2}]Br.

We present upper critical magnetic fields data for kappa(BEDT-TTF)(2)Cu[N(CN)(2)]Br (hereafter k-Br), single crystals with the magnetic field applied either or perpendicular and parallel to the conducting layers. The H-T phase diagram in these single crystals has been deduced from tunnel diode oscillator-based contactless measurements in pulsed magnetic fields up to 56 T for the inter-plane (H parallel to b) and in-plane (H parallel to ac) field directions. Temperature dependence of the magnetic penetration depth in DC fields is also reported. The temperature dependence of the upper critical magnetic field H-c2(T) is not accounted for the Werthamer-Helfand-Hohenberg model for in-plane configuration. For the inter-plane orientation, a significant upward curvature is observed as the temperature decreases, in the range close to T-c, followed by saturation at lower temperatures. Possibly, this upturn is an indication of flux-line lattice melting.

Gasparov, V. A.↗
DOE OSTI2023

HipBone: A performance-portable graphics processing unit-accelerated C++ version of the NekBone benchmark

We present hipBone, an open-source performance-portable proxy application for the Nek5000 (and NekRS) computational fluid dynamics applications. HipBone is a fully GPU-accelerated C++ implementation of the original NekBone CPU proxy application with several novel algorithmic and implementation improvements which optimize its performance on modern fine-grain parallel GPU accelerators. Our optimizations include a conversion to store the degrees of freedom of the problem in assembled form in order to reduce the amount of data moved during the main iteration and a portable implementation of the main Poisson operator kernel. We demonstrate near-roofline performance of the operator kernel on three different modern GPU accelerators from two different vendors. We present a novel algorithm for splitting the application of the Poisson operator on GPUs which aggressively hides MPI communication required for both halo exchange and assembly. Our implementation of nearest-neighbor MPI communication then leverages several different routing algorithms and GPU-Direct RDMA capabilities, when available, which improves scalability of the benchmark. We demonstrate the performance of hipBone on three different clusters housed at Oak Ridge National Laboratory, namely, the Summit supercomputer and the Frontier early-access clusters, Spock and Crusher. Our tests demonstrate both portability across different clusters and very good scaling efficiency, especially on large problems.

Computer Science↗
DOE OSTI2018

Multi-threaded Sparse Matrix Sparse Matrix Multiplication for Many-Core and GPU Architectures.

Sparse Matrix-Matrix multiplication is a key kernel that has applications in several domains such as scientific computing and graph analysis. Several algorithms have been studied in the past for this foundational kernel. In this paper, we develop parallel algorithms for sparse matrix- matrix multiplication with a focus on performance portability across different high performance computing architectures. The performance of these algorithms depend on the data structures used in them. We compare different types of accumulators in these algorithms and demonstrate the performance difference between these data structures. Furthermore, we develop a meta-algorithm, kkSpGEMM, to choose the right algorithm and data structure based on the characteristics of the problem. We show performance comparisons on three architectures and demonstrate the need for the community to develop two phase sparse matrix-matrix multiplication implementations for efficient reuse of the data structures involved.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Alquimia v1.0: a generic interface to biogeochemical codes – a tool for interoperable development, prototyping and benchmarking for multiphysics simulators

Alquimia v1.0 is a generic interface to geochemical solvers that facilitates development of multiphysics simulators by enabling code coupling, prototyping and benchmarking. The interface enforces the function arguments and their types for setting up, solving, serving up output data and carrying out other common auxiliary tasks while providing a set of structures for data transfer between the multiphysics code driving the simulation and the geochemical solver. Alquimia relies on a single-cell approach that permits operator splitting coupling and parallel computation. We describe the implementation in Alquimia of two widely used open-source codes that perform geochemical calculations: PFLOTRAN and CrunchFlow. We then exemplify its use for the implementation and simulation of reactive transport in porous media by two open-source flow and transport simulators: Amanzi and ParFlow. We also demonstrate its use for the simulation of coupled processes in novel multiphysics applications including the effect of multiphase flow on reaction rates at the pore scale with OpenFOAM, the role of complex biogeochemical processes in land surface models such as the E3SM Land Model (ELM) and the impact of surface–subsurface hydrological interactions on hydrogeochemical export from watersheds with the Advanced Terrestrial Simulator (ATS). These applications make it apparent that the availability of a well-defined yet flexible interface has the potential to improve the software development workflow, freeing up resources to focus on advances in process models and mechanistic understanding of coupled problems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2016

Modeling and Analysis Methods for an On-line Enrichment Monitor

The International Atomic Energy Agency (IAEA) has developed an On-Line Enrichment Monitor (OLEM) as one possible component in a new generation of safeguards measures for uranium enrichment plants. The OLEM measures 235U emissions from the UF6 gas flowing through a unit header pipe using NaI(Tl) spectrometers, and corrects for gas density changes using pressure and temperature sensors in order to determine the enrichment of the gas as a function of time. In parallel with the OLEM instrument development, a Virtual OLEM (VOLEM) software tool has been developed that is capable of producing synthetic gamma-ray, pressure, and temperature data representative of a wide range of enrichment plant operating conditions. VOLEM complements instrument development activities and allows the study of OLEM for scenarios that will be difficult or impossible to evaluate empirically. Uses of VOLEM include: investigation of hardware design options; inter-comparison of candidate gamma-ray spectral analysis and enrichment estimation algorithms; uncertainty budget analysis and performance prediction for typical and atypical operational scenarios; and testing of the OLEM data acquisition, analysis and reporting software. This paper describes the technical foundations of VOLEM and illustrates how it can be used. An overview of the nominal instrument design and deployment scenario for OLEM is provided, with emphasis on the key online-assay measurement challenge: accurately determining the portion of the total 235U signal that comes from a background that includes solid uranium deposits on the piping walls. Monte Carlo modeling tools, data analysis algorithms and uncertainty quantification methods are described. VOLEM is then used to quantitatively explore the uncertainty budgets and predicted instrument performance for a plausible range of typical plant operating parameters, and one set of candidate analysis algorithms. Additionally, a series of VOLEM case studies illustrates how an online enrichment monitor might respond in scenarios expected to be atypical, for example weak pressure transients or significant changes in background levels.

Smith, Leon E.↗
DOE OSTI2019

Vectorization of Random Number Generation and Reproducibility of Concurrent Particle Transport Simulation

Efficient random number generation with high quality statistical properties and exact reproducibility of Monte Carlo simulations are important requirements in many areas of computational science. VecRNG is a package providing pseudo-random number generation (pRNG) in the context of a new library VecMath. This library bundles up several general-purpose mathematical utilities, data structures, and algorithms having both SIMD and SIMT (GPUs) support based on VecCore.~Several state-of-the-art RNG algorithms are implemented as kernels supporting parallel generation of random numbers in scalar, vector, and Cuda workflows. In this report, we will present design considerations, implementation details, and computing performance of parallel pRNG engines on both CPU and GPU. Reproducibility of propagating multiple particles in parallel for HEP event simulation is demonstrated, using GeantV-based examples, for both sequential and fine-grain track-level concurrent simulation wor kflows. Strategies for efficient uses of vectorized pRNG and non-overlapping streams of random number sequences in concurrent computing environments is discussed as well.

Jun, S. Y.↗
DOE OSTI2021

Revisiting Huffman Coding: Toward Extreme Performance on Modern GPU Architectures

Today's high-performance computing (HPC) applications are producing vast volumes of data, which are challenging to store and transfer efficiently during the execution, such that data compression is becoming a critical technique to mitigate the storage burden and data movement cost. Huffman coding is arguably the most efficient Entropy coding algorithm in information theory, such that it could be found as a fundamental step in many modern compression algorithms such as DEFLATE. On the other hand, today's HPC applications are more and more relying on the accelerators such as GPU on supercomputers, while Huffman encoding suffers from low throughput on GPUs, resulting in a significant bottleneck in the entire data processing. In this paper, we propose and implement an efficient Huffman encoding approach based on modern GPU architectures, which addresses two key challenges: (1) how to parallelize the entire Huffman encoding algorithm, including codebook construction, and (2) how to fully utilize the high memory-bandwidth feature of modern GPU architectures. The detailed contribution is fourfold. (1) We develop an efficient parallel codebook construction on GPUs that scales effectively with the number of input symbols. (2) We propose a novel reduction based encoding scheme that can efficiently merge the codewords on GPUs. (3) We optimize the overall GPU performance by leveraging the state-of-the-art CUDA APIs such as Cooperative Groups. (4) We evaluate our Huffman encoder thoroughly using six real-world application datasets on two advanced GPUs and compare with our implemented multithreaded Huffman encoder. Experiments show that our solution can improve the encoding throughput by up to 5.0× and 6.8× on NVIDIA RTX 5000 and V100, respectively, over the state-of-the-art GPU Huffman encoder, and by up to 3.3× over the multithread encoder on two 28-core Xeon Platinum 8280 CPUs.

Tian, Jiannan↗
DOE OSTI2021

Integrated transmission electron microscope

An integrated transmission electron microscope comprising multiple electron sources for tuned beams of ultrafast, scanning probe, and parallel illumination in varied beam energies can be alternated within sub-microseconds onto a sample with dynamic ‘transient state’ processes to acquire atomic-scale structural/chemical data with site specificity. The various electron sources and condenser optics enable high-resolution imaging, high-temporal resolution imaging, and chemical imaging, using fast-switching magnets to direct the different electron beams onto a single maneuverable objective pole piece where the sample resides. Such multimodal in situ characterization tools housed in a single microscope have the potential to revolutionize materials science.

Jungjohann, Katherine L.↗
DOE OSTI2022

Systems and methods for powder bed additive manufacturing anomaly detection

Detection and classification of anomalies for powder bed metal additive manufacturing. Anomalies, such as recoater blade impacts, binder deposition issues, spatter generation, and some porosities, are surface-visible at each layer of the building process. A multi-scaled parallel dynamic segmentation convolutional neural network architecture provides additive manufacturing machine and imaging system agnostic pixel-wise semantic segmentation of layer-wise powder bed image data. Learned knowledge is easily transferrable between different additive manufacturing machines. The anomaly detection can be conducted in real-time and provides accurate and generalizable results.

Scime, Luke R.↗
DOE OSTI2016

TU-D-201-03: Results of a Survey On the Implementation of the TG-51 Protocol and Associated Addendum On Reference Dosimetry of External Beams

Purpose: The working group on the review and extension of the TG-51 protocol (WGTG51) collected data from American Association of Physicists in Medicine (AAPM) members with respect to their current TG-51 and associated addendum usage in the interest of considering future protocol addenda and guidance on reference dosimetry best practices. This study reports an overview of this survey on dosimetry of external beams. Methods: Fourteen survey questions were developed by WGTG51 and released in November 2015. The questions collected information on reference dosimetry, beam quality specification, and ancillary calibration equipment. Results: Of the 190 submissions completed worldwide (U.S. 70%), 83% were AAPM members. Of the respondents, 33.5% implemented the TG-51 addendum, with the maximum calibration difference for any photon beam, with respect to the original TG-51 protocol, being <1% for 97.4% of responses. One major finding is that 81.8% of respondents used the same cylindrical ionization chamber for photon and electron dosimetry, implying that many clinics are foregoing the use of parallel-plate chambers. Other evidence suggests equivalent dosimetric results can be obtained with both cylindrical and parallel-plate chambers in electron beams. This, combined with users comfort with cylindrical chambers for electrons will likely impact recommendations put forward in an upcoming electron beam addendum to the TG-51 protocol. Data collected on ancillary equipment showed 58.2% (45.0%) of the thermometers (barometers) in use for beam calibration had NIST traceable calibration certificates, but 48.4% (42.7%) were never recalibrated. Conclusion: This survey provides a snapshot of TG-51 external beam reference dosimetry practice in radiotherapy centers. Findings demonstrate the rapid take-up of the TG-51 photon beam addendum and raise issues for the WGTG51 to focus on going forward, including guidelines on ancillary equipment and the choice of chamber for electron beam dosimetry.

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2024

Boosting RDataFrame performance with transparent bulk event processing

RDataFrame is ROOT’s high-level interface for Python and C++ data analysis. Since it first became available, RDataFrame adoption has grown steadily and it is now poised to be a major component of analysis software pipelines for LHC Run 3 and beyond. Thanks to its design inspired by declarative programming principles, RDataFrame enables the development of highperformance, highly parallel analyses without requiring expert knowledge of multi-threading and I/O: user logic is expressed in terms of self-contained, small computation kernels tied together by a high-level API. This design completely decouples analysis logic from its actual execution, and opens several interesting avenues for workflow optimization. In particular, in this work we explore the benefits of moving internal data processing from an event-by-event to a bulkby-bulk loop. This refactoring dramatically reduces the framework’s runtime overheads; in collaboration with the I/O layer it improves data access patterns; it exposes information that optimizing compilers might use to auto-vectorize the invocation of user-defined computations; finally, while existing user-facing interfaces remain unaffected, it becomes possible to additionally offer interfaces that explicitly expose bulks of events, useful e.g. for the injection of GPU kernels into the analysis workflow. In order to inform similar future R&D, design challenges will be presented, as well as an investigation of the relevant timememory trade-off backed by novel performance benchmarks.

Guiraud, Enrico↗
DOE OSTI2016

IAEA CRP on HTGR Uncertainties in Modeling: Assessment of Phase I Lattice to Core Model Uncertainties

Best-estimate plus uncertainty analysis of reactors is replacing the traditional conservative (stacked uncertainty) method for safety and licensing analysis. To facilitate uncertainty analysis applications, a comprehensive approach and methodology must be developed and applied. High temperature gas cooled reactors (HTGRs) have several features that require techniques not used in light-water reactor analysis (e.g., coated-particle design and large graphite quantities at high temperatures). The International Atomic Energy Agency has therefore launched the Coordinated Research Project on HTGR Uncertainty Analysis in Modeling to study uncertainty propagation in the HTGR analysis chain. The benchmark problem defined for the prismatic design is represented by the General Atomics Modular HTGR 350. The main focus of this report is the compilation and discussion of the results obtained for various permutations of Exercise I 2c and the use of the cross section data in Exercise II 1a of the prismatic benchmark, which is defined as the last and first steps of the lattice and core simulation phases, respectively. The report summarizes the Idaho National Laboratory (INL) best estimate results obtained for Exercise I 2a (fresh single-fuel block), Exercise I 2b (depleted single-fuel block), and Exercise I 2c (super cell) in addition to the first results of an investigation into the cross section generation effects for the super-cell problem. The two dimensional deterministic code known as the New ESC based Weighting Transport (NEWT) included in the Standardized Computer Analyses for Licensing Evaluation (SCALE) 6.1.2 package was used for the cross section evaluation, and the results obtained were compared to the three dimensional stochastic SCALE module KENO VI. The NEWT cross section libraries were generated for several permutations of the current benchmark super-cell geometry and were then provided as input to the Phase II core calculation of the stand alone neutronics Exercise II 1a. The steady state core calculations were simulated with the INL coupled-code system known as the Parallel and Highly Innovative Simulation for INL Code System (PHISICS) and the system thermal-hydraulics code known as the Reactor Excursion and Leak Analysis Program (RELAP) 5 3D using the nuclear data libraries previously generated with NEWT. It was observed that significant differences in terms of multiplication factor and neutron flux exist between the various permutations of the Phase I super-cell lattice calculations. The use of these cross section libraries only leads to minor changes in the Phase II core simulation results for fresh fuel but shows significantly larger discrepancies for spent fuel cores. Furthermore, large incongruities were found between the SCALE NEWT and KENO VI results for the super cells, and while some trends could be identified, a final conclusion on this issue could not yet be reached. This report will be revised in mid 2016 with more detailed analyses of the super-cell problems and their effects on the core models, using the latest version of SCALE (6.2). The super-cell models seem to show substantial improvements in terms of neutron flux as compared to single-block models, particularly at thermal energies.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗