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 1,369 records · Page 76

Distributed deformation and block rotation in 3D

The authors address how block rotation and complex distributed deformation in the Earth's shallow crust may be explained within a stationary regional stress field. Distributed deformation is characterized by domains of sub-parallel fault-bounded blocks. In response to the contemporaneous activity of neighboring domains some domains rotate, as suggested by both structural and paleomagnetic evidence. Rotations within domains are achieved through the contemporaneous slip and rotation of the faults and of the blocks they bound. Thus, in regions of distributed deformation, faults must remain active in spite of their poor orientation in the stress field. The authors developed a model that tracks the orientation of blocks and their bounding faults during rotation in a 3D stress field. In the model, the effective stress magnitudes of the principal stresses (sigma sub 1, sigma sub 2, and sigma sub 3) are controlled by the orientation of fault sets in each domain. Therefore, adjacent fault sets with differing orientations may be active and may display differing faulting styles, and a given set of faults may change its style of motion as it rotates within a stationary stress regime. The style of faulting predicted by the model depends on a dimensionless parameter phi = (sigma sub 2 - sigma sub 3)/(sigma sub 1 - sigma sub 3). Thus, the authors present a model for complex distributed deformation and complex offset history requiring neither geographical nor temporal changes in the stress regime. They apply the model to the Western Transverse Range domain of southern California. There, it is mechanically feasible for blocks and faults to have experienced up to 75 degrees of clockwise rotation in a phi = 0.1 strike-slip stress regime. The results of the model suggest that this domain may first have accommodated deformation along preexisting NNE-SSW faults, reactivated as normal faults. After rotation, these same faults became strike-slip in nature.

Scotti, Oona↗

Gyrocenter-gauge kinetic algorithm for high frequency waves in magnetized plasmas

A kinetic simulation algorithm for high-frequency electromagnetic waves has been developed based on the gyrocenter-gauge kinetic theory. The magnetized plasma system is simulated in the gyrocenter coordinate system. The gyrocenter distribution function F is sampled on the gyrocenter, parallel velocity, and magnetic moment coordinates. The gyrocenter-gauge function S is sampled on the Kruskal rings and shares the first five coordinates with F. The moment integral of pullback transformation is directly calculated using the Monte Carlo method and an explicit difference scheme for Maxwell's equations in terms of potentials is adopted. The new algorithm has been successfully applied to the simulation studies of high frequency extraordinary wave, electron Bernstein wave, and the mode conversion process between the extraordinary wave and the electron Bernstein wave in inhomogeneous plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Parallel sorting algorithm classification: is manual instrumentation necessary?

Understanding parallel algorithms is crucial for accelerating scientific simulations on complex, distributed memory, high-performance computers. Modern algorithm classification approaches learn semantics directly from source code to differentiate between algorithms, however, accessing source code is not always possible. We can learn about parallel algorithms from observing their performance, as programs running the same algorithms and using the same hardware should exhibit similar performance characteristics. We present an approach to learn algorithm classes from parallel performance data directly in order to classify algorithms without access to the source code. We extend previous work to enable classifying parallel sorting algorithms using automatic instrumentation instead of requiring manual region annotations in the source code. In this work, we design and demonstrate a study for classification of parallel sorting algorithms using parallel performance data collected from automatic instrumentation, and evaluate the performance of our new methodology on classification. We leverage Caliper to collect the performance data, Thicket for our exploratory data analysis (EDA), and PyTorch and Scikit-learn to evaluate the effectiveness of random forests, support vector machines (SVMs), decision trees, neural networks, and logistic regressions on parallel performance data. Additionally, we study noise in parallel performance data, whether the removal of noise and pre-processing of the data is necessary to accurately classify parallel sorting algorithms, and determine the effectiveness of features created from performance data. In conclusion, we demonstrate classification accuracy for these five different models of up to 97.7% across four different parallel algorithm classes.

Algorithm Classification↗

A prospectus on kinetic heliophysics

Under the low density and high temperature conditions typical of heliospheric plasmas, the macroscopic evolution of the heliosphere is strongly affected by the kinetic plasma physics governing fundamental microphysical mechanisms. Kinetic turbulence, collisionless magnetic reconnection, particle acceleration, and kinetic instabilities are four poorly understood, grand-challenge problems that lie at the new frontier of kinetic heliophysics. The increasing availability of high cadence and high phase-space resolution measurements of particle velocity distributions by current and upcoming spacecraft missions and of massively parallel nonlinear kinetic simulations of weakly collisional heliospheric plasmas provides the opportunity to transform our understanding of these kinetic mechanisms through the full utilization of the information contained in the particle velocity distributions. Several major considerations for future investigations of kinetic heliophysics are examined. Furthermore, turbulent dissipation followed by particle heating is highlighted as an inherently two-step process in weakly collisional plasmas, distinct from the more familiar case in fluid theory.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Automatic Data Distribution for CFD Applications on Structured Grids

Development of HPF versions of NPB and ARC3D showed that HPF has potential to be a high level language for parallelization of CFD applications. The use of HPF requires an intimate knowledge of the applications and a detailed analysis of data affinity, data movement and data granularity. Since HPF hides data movement from the user even with this knowledge it is easy to overlook pieces of the code causing low performance of the application. In order to simplify and accelerate the task of developing HPF versions of existing CFD applications we have designed and partially implemented ADAPT (Automatic Data Distribution and Placement Tool). The ADAPT analyzes a CFD application working on a single structured grid and generates HPF TEMPLATE, (RE)DISTRIBUTION, ALIGNMENT and INDEPENDENT directives. The directives can be generated on the nest level, subroutine level, application level or inter application level. ADAPT is designed to annotate existing CFD FORTRAN application performing computations on single or multiple grids. On each grid the application can considered as a sequence of operators each applied to a set of variables defined in a particular grid domain. The operators can be classified as implicit, having data dependences, and explicit, without data dependences. In order to parallelize an explicit operator it is sufficient to create a template for the domain of the operator, align arrays used in the operator with the template, distribute the template, and declare the loops over the distributed dimensions as INDEPENDENT. In order to parallelize an implicit operator, the distribution of the operator's domain should be consistent with the operator's dependences. Any dependence between sections distributed on different processors would preclude parallelization if compiler does not have an ability to pipeline computations. If a data distribution is "orthogonal" to the dependences of an implicit operator then the loop which implements the operator can be declared as INDEPENDENT.

Frumkin, Michael↗

Technical Note: Range verification system using edge detection method for a scintillator and a CCD camera system

Purpose: Three-dimensional irradiation with a scanned carbon-ion beam has been performed from 2011 at the authors’ facility. The authors have developed the rotating-gantry equipped with the scanning irradiation system. The number of combinations of beam properties to measure for the commissioning is more than 7200, i.e., 201 energy steps, 3 intensities, and 12 gantry angles. To compress the commissioning time, quick and simple range verification system is required. In this work, the authors develop a quick range verification system using scintillator and charge-coupled device (CCD) camera and estimate the accuracy of the range verification. Methods: A cylindrical plastic scintillator block and a CCD camera were installed on the black box. The optical spatial resolution of the system is 0.2 mm/pixel. The camera control system was connected and communicates with the measurement system that is part of the scanning system. The range was determined by image processing. Reference range for each energy beam was determined by a difference of Gaussian (DOG) method and the 80% of distal dose of the depth-dose distribution that were measured by a large parallel-plate ionization chamber. The authors compared a threshold method and a DOG method. Results: The authors found that the edge detection method (i.e., the DOG method) is best for the range detection. The accuracy of range detection using this system is within 0.2 mm, and the reproducibility of the same energy measurement is within 0.1 mm without setup error. Conclusions: The results of this study demonstrate that the authors’ range check system is capable of quick and easy range verification with sufficient accuracy.

60 APPLIED LIFE SCIENCES↗

Alfven waves in dusty plasmas with plasma particles described by anisotropic kappa distributions

We utilize a kinetic description to study the dispersion relation of Alfven waves propagating parallelly to the ambient magnetic field in a dusty plasma, taking into account the fluctuation of the charge of the dust particles, which is due to inelastic collisions with electrons and ions. We consider a plasma in which the velocity distribution functions of the plasma particles are modelled as anisotropic kappa distributions, study the dispersion relation for several combinations of the parameters {kappa}{sub Parallel-To} and {kappa}{sub Up-Tack }, and emphasize the effect of the anisotropy of the distributions on the mode coupling which occurs in a dusty plasma, between waves in the branch of circularly polarized waves and waves in the whistler branch.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Explicit spatial scattering for load balancing in conservatively synchronized parallel discrete-event simulations

We re-examine the problem of load balancing in conservatively synchronized parallel, discrete-event simulations executed on high-performance computing clusters, focusing on simulations where computational and messaging load tend to be spatially clustered. Such domains are frequently characterized by the presence of geographic 'hot-spots' - regions that generate significantly more simulation events than others. Examples of such domains include simulation of urban regions, transportation networks and networks where interaction between entities is often constrained by physical proximity. Noting that in conservatively synchronized parallel simulations, the speed of execution of the simulation is determined by the slowest (i.e most heavily loaded) simulation process, we study different partitioning strategies in achieving equitable processor-load distribution in domains with spatially clustered load. In particular, we study the effectiveness of partitioning via spatial scattering to achieve optimal load balance. In this partitioning technique, nearby entities are explicitly assigned to different processors, thereby scattering the load across the cluster. This is motivated by two observations, namely, (i) since load is spatially clustered, spatial scattering should, intuitively, spread the load across the compute cluster, and (ii) in parallel simulations, equitable distribution of CPU load is a greater determinant of execution speed than message passing overhead. Through large-scale simulation experiments - both of abstracted and real simulation models - we observe that scatter partitioning, even with its greatly increased messaging overhead, significantly outperforms more conventional spatial partitioning techniques that seek to reduce messaging overhead. Further, even if hot-spots change over the course of the simulation, if the underlying feature of spatial clustering is retained, load continues to be balanced with spatial scattering leading us to the observation that spatial scattering can often obviate the need for dynamic load balancing.

97 MATHEMATICS AND COMPUTING↗

Management of the interplay effect when using dynamic MLC sequences to treat moving targets

Interplay between organ motion and leaf motion has been shown to generally have a small dosimetric impact for most clinical intensity-modulated radiation therapy treatments. However, it has also been shown that for some MLC sequences there can be large daily variations in the delivered dose, depending on details of patient motion or the number of fractions. This study investigates guidelines for dynamic MLC sequences that will keep daily dose variations due to the interplay between organ motion and leaf motion within 10%. Dose distributions for a range of MLC separations (0.2-5.0 cm) and displacements between adjacent MLCs (0-1.5 cm) were exported from ECLIPSE to purpose-written software, which simulated the dose distribution delivered to a moving target. Target motion parallel and perpendicular to the MLC motion was investigated for a range of amplitudes (0.5-4.0 cm), periods (1.5-10 s), and MLC speeds (0.1-3.0 cm/s) with target motions modeled as sin{sup 6}. Results were confirmed experimentally by measuring the dose delivered to an ion chamber array in a moving phantom for different MLC sequences. The simulation results were used to identify MLC sequences that kept dose variations within 10% compared to the dose delivered with no motion. The maximum allowable MLC speed, when target motion is parallel to the MLC motion, was found to be a simple function of target period and MLC separation. When the target motion is perpendicular to MLC motion, the maximum allowable MLC speed can be described as a function of MLC separation and the displacement of adjacent MLCs. These guidelines were successfully applied to two-dimensional motion, and a simple program was written to import MLC sequence files and evaluate whether the maximum daily dose discrepancy caused by the interplay effect will be larger than 10%. This software was experimentally evaluated, and found to conservatively predict whether a given MLC sequence could give large daily dose discrepancies.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

SHARP: a distributed GPU-based ptychographic solver

Ever brighter light sources, fast parallel detectors and advances in phase retrieval methods have made ptychography a practical and popular imaging technique. Compared to previous techniques, ptychography provides superior robustness and resolution at the expense of more advanced and time-consuming data analysis. Therefore, by taking advantage of massively parallel architectures, high-throughput processing can expedite this analysis and provide microscopists with immediate feedback. These advances allow real-time imaging at wavelength-limited resolution, coupled with a large field of view. This article describes a set of algorithmic and computational methodologies used at the Advanced Light Source and US Department of Energy light sources. These are packaged as a CUDA-based software environment named SHARP, aimed at providing state-of-the-art high-throughput ptychography reconstructions for the coming era of diffraction-limited light sources.

47 OTHER INSTRUMENTATION↗

Efficient algorithms and implementations of entropy-based moment closures for rarefied gases

We present efficient algorithms and implementations of the 35-moment system equipped with the maximum-entropy closure in the context of rarefied gases. While closures based on the principle of entropy maximization have been shown to yield very promising results for moderately rarefied gas flows, the computational cost of these closures is in general much higher than for closure theories with explicit closed-form expressions of the closing fluxes, such as Grad's classical closure. Following a similar approach as Garrett et al. (2015) , we investigate efficient implementations of the computationally expensive numerical quadrature method used for the moment evaluations of the maximum-entropy distribution by exploiting its inherent fine-grained parallelism with the parallelism offered by multi-core processors and graphics cards. We show that using a single graphics card as an accelerator allows speed-ups of two orders of magnitude when compared to a serial CPU implementation. To accelerate the time-to-solution for steady-state problems, we propose a new semi-implicit time discretization scheme. The resulting nonlinear system of equations is solved with a Newton type method in the Lagrange multipliers of the dual optimization problem in order to reduce the computational cost. Additionally, fully explicit time-stepping schemes of first and second order accuracy are presented. We investigate the accuracy and efficiency of the numerical schemes for several numerical test cases, including a steady-state shock-structure problem.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Addressing Two-Phase Flow Maldistribution in Microchannel Heat and Mass Exchangers

Multiphase flow phenomena in single micro and minichannels have been widely studied. Characteristics of two-phase flow through a large array of microchannels are investigated here. An air–water mixture is used to represent the two phases flowing through a microchannel array representative of those employed in practical applications. Flow distribution of the air and water flow across 52 parallel microchannels of 0.4 mm hydraulic diameter is visually investigated using high-speed photography. Two microchannel configurations are studied and compared, with mixing features incorporated into the second configuration. Slug and annular flow regimes are observed in the channels. Void fractions and interfacial areas are calculated for each channel from these observations. The flow distribution is tracked at various lengths along the microchannel array sheets. Statistical distributions of void fraction and interfacial area along the microchannel array are measured. The design with mixing features yields improved flow distribution. Void fraction and interfacial area change along the length of the second configuration, indicating a change in fluid distribution among the channels. The void fraction and interfacial area results are used to predict the performance of different microchannel array configurations for heat and mass transfer applications. Results from this study can help inform the design of compact thermal-fluid energy systems.

Engineering↗

Evidence for parallel electric field particle acceleration in the dayside auroral oval

Electron and ion energy spectra and electron pitch angle distributions are presented for two sounding rocket flights in the dayside auroral zone. At times, effects of dc electric fields parallel to the magnetic field are evident in that: (1) within precipitation features, protons are decelerated by an amount of energy consistent with that which electrons gain and (2) electrons are sometimes aligned to within 3 deg (full width at half maximum) of the magnetic field. A maximum altitude for the accelerating region of several thousand kilometers is deduced from the narrow width of the pitch angle distribution and also from time-of-flight delays between the observation of accelerated electrons and decelerated protons.

Torbert, R. B.↗

Thermal response of a series- and parallel-connected solar energy storage to multi-day charge sequences

The thermal response of a multi-tank thermal storage was studied under variable charge conditions. Tests were conducted on an experimental apparatus that simulated the thermal charging of the storage system by a solar collector over predetermined (prescribed) daylong periods. The storage was assembled from three standard 270 L hot-water storage tanks each charged through coupled, side-arm, natural convection heat exchangers which were connected in either a series- or parallel-flow configuration. Both energy storage rates and tank temperature profiles were experimentally measured during charge periods representative of two consecutive clear days or combinations of a clear and overcast day. During this time, no draw-offs were conducted. Of particular interest was the effect of rising and falling charge-loop temperatures and collector-loop flow rate on storage tank stratification levels. Results of this study show that the series-connected thermal storage reached high levels of temperature stratification in the storage tanks during periods of rising charge temperatures and also limited destratification during periods of falling charge temperature. This feature is a consequence of the series-connected configuration that allowed sequential stratification to occur in the component tanks and energy to be distributed according to temperature level. This effect was not observed in the parallel charge configuration. A further aspect of the study investigated the effect of increasing charge-loop flow rate on the temperature distribution within the series-connected storage and showed that, at high flow rates, the temperature distributions were found to be similar to those obtained during parallel charging. A disadvantage of both the high-flow series-connected and parallel-connected multi-tank storage is that falling charge-loop temperatures, which normally occur in the afternoon, tend to mix and destratify the storage tanks. (author)

14 SOLAR ENERGY↗

Genarris 2.0: A Random Structure Generator for Molecular Crystals

Genarris is an open source Python package for generating random molecular crystal structures with physical constraints for seeding crystal structure prediction algorithms and training machine learning models. Here we present a new version of the code, containing several major improvements. A MPI-based parallelization scheme has been implemented, which facilitates the seamless sequential execution of user-defined workflows. A new method for estimating the unit cell volume based on the single molecule structure has been developed using a machine-learned model trained on experimental structures. A new algorithm has been implemented for generating crystal structures with molecules occupying special Wyckoff positions. A new hierarchical structure check procedure has been developed to detect unphysical close contacts efficiently and accurately. New intermolecular distance settings have been implemented for strong hydrogen bonds. To demonstrate these new features, we study two specific cases: benzene and glycine. Genarris finds the experimental structures of the two polymorphs of benzene and the three polymorphs of glycine. Program summary Program Title: Genarris 2.0 Program Files doi: http://dx.doi.org/10.17632/grx6mz4pjn.1 Licensing provisions: BSD-3 Clause Programming language: Python, C External routines/libraries: Spglib, ASE, pymatgen, SciPy, mpi4py, scikit-learn, PyTorch, FHI-aims. Nature of problem: Molecular crystal structure prediction. Solution method: Genarris 2.0 generates molecular crystal structures over the 230 space groups, on general and special Wyckoff positions, using physical constraints. Down-sampling of the generated structures may be performed subsequently, based on molecular crystal packing descriptors and an unsupervised machine learning algorithm. Lastly, ab initio structure relaxation may be performed for the final pool. Depending on the user-defined workflow implemented, Genarris may be used to generate diverse molecular crystal datasets to seed evolutionary algorithms or to train machine learning algorithms or as a standalone crystal structure prediction method. Restrictions: For crystal structure generation, the molecule of interest must be semi-rigid with no bond rotational degrees of freedom. Unusual features: Genarris 2.0 is a highly distributed program, making use of MPI for Python parallelization. The user has the ability to design and implement workflows by executing a user-defined list of procedures. Genarris 2.0 offers new features including a machine learning model for estimating the molecular volume in the solid state from the single molecule structure, structure generation in special Wyckoff positions of space groups, hierarchical structure checks including rigorous treatment of non-orthogonal structures, and clustering and down-selection workflows combining first principles simulations with machine learning. (C) 2020 Elsevier B.V. All rights reserved.

Crystal structure prediction↗

LANDSAT-D investigations in snow hydrology

Work undertaken during the contract and its results are described. Many of the results from this investigation are available in journal or conference proceedings literature - published, accepted for publication, or submitted for publication. For these the reference and the abstract are given. Those results that have not yet been submitted separately for publication are described in detail. Accomplishments during the contract period are summarized as follows: (1) analysis of the snow reflectance characteristics of the LANDSAT Thematic Mapper, including spectral suitability, dynamic range, and spectral resolution; (2) development of a variety of atmospheric models for use with LANDSAT Thematic Mapper data. These include a simple but fast two-stream approximation for inhomogeneous atmospheres over irregular surfaces, and a doubling model for calculation of the angular distribution of spectral radiance at any level in an plane-parallel atmosphere; (3) incorporation of digital elevation data into the atmospheric models and into the analysis of the satellite data; and (4) textural analysis of the spatial distribution of snow cover.

Dozier, J.↗

Parallel fast gauss transform

We present fast adaptive parallel algorithms to compute the sum of N Gaussians at N points. Direct sequential computation of this sum would take O(N{sup 2}) time. The parallel time complexity estimates for our algorithms are O(N/n{sub p}) for uniform point distributions and O( (N/n{sub p}) log (N/n{sub p}) + n{sub p}log n{sub p}) for non-uniform distributions using n{sub p} CPUs. We incorporate a plane-wave representation of the Gaussian kernel which permits 'diagonal translation'. We use parallel octrees and a new scheme for translating the plane-waves to efficiently handle non-uniform distributions. Computing the transform to six-digit accuracy at 120 billion points took approximately 140 seconds using 4096 cores on the Jaguar supercomputer. Our implementation is 'kernel-independent' and can handle other 'Gaussian-type' kernels even when explicit analytic expression for the kernel is not known. These algorithms form a new class of core computational machinery for solving parabolic PDEs on massively parallel architectures.

99 GENERAL AND MISCELLANEOUS↗

Some fast elliptic solvers on parallel architectures and their complexities

The discretization of separable elliptic partial differential equations leads to linear systems with special block tridiagonal matrices. Several methods are known to solve these systems, the most general of which is the Block Cyclic Reduction (BCR) algorithm which handles equations with nonconstant coefficients. A method was recently proposed to parallelize and vectorize BCR. In this paper, the mapping of BCR on distributed memory architectures is discussed, and its complexity is compared with that of other approaches including the Alternating-Direction method. A fast parallel solver is also described, based on an explicit formula for the solution, which has parallel computational compelxity lower than that of parallel BCR.

Gallopoulos, E.↗