Search NASA⌕ Search

SEARCH · Search NASA

Results for “data parallel algorithms”

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 919 records · Page 51
All sourcesNASA NTRSDOE OSTI
NASA NTRS1999

Experimental Investigation of Pool Boiling Heat Transfer Enhancement in Microgravity in the Presence of Electric Fields

In boiling high heat fluxes are possible driven by relatively small temperature differences, which make its use increasingly attractive in aerospace applications. The objective of the research is to develop ways to overcome specific problems associated with boiling in the low gravity environment by substituting the buoyancy force with the electric force to enhance bubble removal from the heated surface. Previous studies indicate that in terrestrial applications nucleate boiling heat transfer can be increased by a factor of 50, as compared to values obtained for the same system without electric fields. The goal of our research is to experimentally explore the mechanisms responsible for EHD heat transfer enhancement in boiling in low gravity conditions, by visualizing the temperature distributions in the vicinity of the heated surface and around the bubble during boiling using real-time holographic interferometry (HI) combined with high-speed cinematography. In the first phase of the project the influence of the electric field on a single bubble is investigated. Pool boiling is simulated by injecting a single bubble through a nozzle into the subcooled liquid or into the thermal boundary layer developed along the flat heater surface. Since the exact location of bubble formation is known, the optical equipment can be aligned and focused accurately, which is an essential requirement for precision measurements of bubble shape, size and deformation, as well as the visualization of temperature fields by HI. The size of the bubble and the frequency of bubble departure can be controlled by suitable selection of nozzle diameter and mass flow rate of vapor. In this approach effects due to the presence of the electric field can be separated from effects caused by the temperature gradients in the thermal boundary layer. The influence of the thermal boundary layer can be investigated after activating the heater at a later stage of the research. For the visualization experiments a test cell was developed. All four vertical walls of the test cell are transparent, and they allow transillumination with laser light for visualization experiments by HI. The bottom electrode is a copper cylinder, which is electrically grounded. The copper block is heated with a resistive heater and it is equipped with 6 thermocouples that provide reference temperatures for the measurements with HI. The top electrode is a mesh electrode. Bubbles are injected with a syringe into the test cell through the bottom electrode. The working fluids presently used in the interferometric visualization experiments, water and PF 5052, satisfy requirements regarding thermophysical, optical and electrical properties. A 30kV power supply equipped with a voltmeter allows to apply the electric field to the electrodes during the experiments. The magnitude of the applied voltage can be adjusted either manually or through the LabVIEW data acquisition and control system connected to a PC. Temperatures of the heated block are recorded using type-T thermocouples, whose output is read by a data acquisition system. Images of the bubbles are recorded with 35mm photographic and 16mm high-speed cameras, scanned and analyzed using various software packages. Visualized temperature fields HI allows the visualization of temperature fields in the vicinity of bubbles during boiling in the form of fringes. Typical visualized temperature distributions around the air bubbles injected into the thermal boundary layer in PF5052 are shown. The temperature of the heated surface is 35 C. The temperature difference for a pair of fringes is approximately 0.05 C. The heat flux applied to the bottom surface is moderate, and the fringe patterns are regular. In the image a bubble penetrating the thermal boundary layer is visible. Because of the axial symmetry of the problem, simplified reconstruction techniques can be applied to recover the temperature field. The thermal plume developing above the heated surface for more intensive heating is shown. The temperature distribution in the liquid is clearly 3D, and tomographic techniques have to be applied to recover the temperature distribution in such a physical situation. A sequence of interferometric images showing the temperature distribution around the rising bubble, recorded with a high-speed camera is shown. Again, the temperature distribution is 3D, and a more complex approach to the evaluation, the tomographic reconstruction has to be taken. Measurement of the temperature distribution from the fringe pattern temperature distributions that yield important information regarding heat transfer are determined. Two algorithms that allow the quantitative evaluation of interferometric fringe patterns and the reconstruction of temperature fields during boiling have been developed at the Heat Transfer Laboratory of the Johns Hopkins University. In the first algorithm the bubble is assumed to be axially symmetrical, which significantly reduces the computational effort for quantifying the temperature distribution around the bubble. For this purpose the thermal boundary layer around the bubble is divided into equidistant concentric shells, and the refractive index is assumed to be constant in each of the shells. Since large temperature gradients are expected in the vicinity of the bubble during boiling, the deflection of the light beam cannot be neglected in boiling experiments. Since the exit angle of the light beam is known, this allows to account for the deflections and phase shifts outside the boundary layer (in the bulk fluid and in the windows of the test cell). Three dimensional temperature distributions in the vicinity of the bubble are reconstructed using tomographic techniques. In tomography, the measurement volume is sliced into 2D planes. In the present study these planes are parallel to the heated surface. The objective is to determine the values of the field parameter of interest in form of the field function in these 2D planes. The field parameter is the change of the refractive index of the liquid in the measurement volume caused by temperature changes. By superimposing data for many 2D planes recorded at the same time instant, the 3D temperature distribution in the measurement volume is recovered.

Herman, Cila↗
DOE OSTI2025

Communication Lower Bounds and Optimal Algorithms for Symmetric Matrix Computations

In this article, we focus on the communication costs of three symmetric matrix computations: (i) multiplying a matrix with its transpose, known as a symmetric rank-k update (SYRK) (ii) adding the result of the multiplication of a matrix with the transpose of another matrix and the transpose of that result, known as a symmetric rank-2k update (SYR2K) (iii) performing matrix multiplication with a symmetric input matrix (SYMM). All three computations appear in the Level 3 Basic Linear Algebra Subroutines (BLAS) and have wide use in applications involving symmetric matrices. We establish communication lower bounds for these kernels using sequential and distributed-memory parallel computational models, and we show that our bounds are tight by presenting communication-optimal algorithms for each setting. Our lower bound proofs rely on applying a geometric inequality for symmetric computations and analytically solving constrained nonlinear optimization problems. As a result, the symmetric matrix and its corresponding computations are accessed and performed according to a triangular block partitioning scheme in the optimal algorithms.

Al Daas, Hussam [Rutherford Appleton Laboratory, D↗
DOE OSTI2017

Milestone Completion Report WBS 1.3.5.05 ECP/VTK-m FY17Q2 [MS-17/01] Better Dynamic Types Design SDA05-1

The FY17Q2 milestone of the ECP/VTK-m project, which is the first milestone, includes the completion of design documents for the introduction of virtual methods into the VTK-m framework. Specifically, the ability from within the code of a device (e.g. GPU or Xeon Phi) to jump to a virtual method specified at run time. This change will enable us to drastically reduce the compile time and the executable code size for the VTK-m library. Our first design introduced the idea of adding virtual functions to classes that are used during algorithm execution. (Virtual methods were previously banned from the so called execution environment.) The design was straightforward. VTK-m already has the generic concepts of an “array handle” that provides a uniform interface to memory of different structures and an “array portal” that provides generic access to said memory. These array handles and portals use C++ templating to adjust them to different memory structures. This composition provides a powerful ability to adapt to data sources, but requires knowing static types. The proposed design creates a template specialization of an array portal that decorates another array handle while hiding its type. In this way we can wrap any type of static array handle and then feed it to a single compiled instance of a function. The second design focused on the mechanics of implementing virtual methods on parallel devices with a focus on CUDA. Our initial experiments on CUDA showed a very large overhead for using virtual C++ classes with virtual methods, the standard approach. Instead, we are using an alternate method provided by C that uses function pointers. With the completion of this milestone, we are able to move to the implementation of objects with virtual (like) methods. The upshot will be much faster compile times and much smaller library/executable sizes.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1997

Parallel Computation of Unsteady Flows on a Network of Workstations

Parallel computation of unsteady flows requires significant computational resources. The utilization of a network of workstations seems an efficient solution to the problem where large problems can be treated at a reasonable cost. This approach requires the solution of several problems: 1) the partitioning and distribution of the problem over a network of workstation, 2) efficient communication tools, 3) managing the system efficiently for a given problem. Of course, there is the question of the efficiency of any given numerical algorithm to such a computing system. NPARC code was chosen as a sample for the application. For the explicit version of the NPARC code both two- and three-dimensional problems were studied. Again both steady and unsteady problems were investigated. The issues studied as a part of the research program were: 1) how to distribute the data between the workstations, 2) how to compute and how to communicate at each node efficiently, 3) how to balance the load distribution. In the following, a summary of these activities is presented. Details of the work have been presented and published as referenced.

Source record↗
DOE OSTI2026

BM3DORNL

BM3DORNL is a high-performance, open-source library for removing streak and ring artifacts from computed-tomography (CT) data, developed for neutron imaging at Oak Ridge National Laboratory's Spallation Neutron Source (VENUS beamline) and applicable to X-ray CT as well. Ring artifacts — concentric rings in reconstructed slices caused by detector pixel-to-pixel response non-uniformities — appear as vertical streaks in the sinogram and degrade both image quality and quantitative analysis. BM3DORNL operates in the sinogram domain using an adaptation of the BM3D (block-matching and 3D collaborative filtering) algorithm (Dabov et al., 2007). It provides a dedicated streak-removal mode, a true multi-scale BM3D variant (after Mäkinen et al., 2021) that suppresses wide streaks single-scale methods miss, and an alternative Fourier–SVD method (~2.6× faster) combining FFT-based energy detection with rank-1 SVD. The computationally intensive core is implemented in Rust with parallel (Rayon) block matching, integral-image pre-screening, and optimized transforms, and is exposed through a simple Python API (with an optional GUI) so it integrates directly into existing tomography reconstruction pipelines. It processes both 2D sinograms and 3D sinogram stacks, is pip-installable for Linux and macOS, and is documented at https://bm3dornl.readthedocs.io.

Zhang, Chen [Oak Ridge National Laboratory (ORNL),↗
NASA NTRS1990

Adaptive pattern recognition by mini-max neural networks as a part of an intelligent processor

In this decade and progressing into 21st Century, NASA will have missions including Space Station and the Earth related Planet Sciences. To support these missions, a high degree of sophistication in machine automation and an increasing amount of data processing throughput rate are necessary. Meeting these challenges requires intelligent machines, designed to support the necessary automations in a remote space and hazardous environment. There are two approaches to designing these intelligent machines. One of these is the knowledge-based expert system approach, namely AI. The other is a non-rule approach based on parallel and distributed computing for adaptive fault-tolerances, namely Neural or Natural Intelligence (NI). The union of AI and NI is the solution to the problem stated above. The NI segment of this unit extracts features automatically by applying Cauchy simulated annealing to a mini-max cost energy function. The feature discovered by NI can then be passed to the AI system for future processing, and vice versa. This passing increases reliability, for AI can follow the NI formulated algorithm exactly, and can provide the context knowledge base as the constraints of neurocomputing. The mini-max cost function that solves the unknown feature can furthermore give us a top-down architectural design of neural networks by means of Taylor series expansion of the cost function. A typical mini-max cost function consists of the sample variance of each class in the numerator, and separation of the center of each class in the denominator. Thus, when the total cost energy is minimized, the conflicting goals of intraclass clustering and interclass segregation are achieved simultaneously.

Szu, Harold H.↗
NASA NTRS2009

A Survey of New Trends in Symbolic Execution for Software Testing and Analysis

Symbolic execution is a well-known program analysis technique which represents values of program inputs with symbolic values instead of concrete (initialized) data and executes the program by manipulating program expressions involving the symbolic values. Symbolic execution has been proposed over three decades ago but recently it has found renewed interest in the research community, due in part to the progress in decision procedures, availability of powerful computers and new algorithmic developments. We provide a survey of some of the new research trends in symbolic execution, with particular emphasis on applications to test generation and program analysis. We first describe an approach that handles complex programming constructs such as input data structures, arrays, as well as multi-threading. We follow with a discussion of abstraction techniques that can be used to limit the (possibly infinite) number of symbolic configurations that need to be analyzed for the symbolic execution of looping programs. Furthermore, we describe recent hybrid techniques that combine concrete and symbolic execution to overcome some of the inherent limitations of symbolic execution, such as handling native code or availability of decision procedures for the application domain. Finally, we give a short survey of interesting new applications, such as predictive testing, invariant inference, program repair, analysis of parallel numerical programs and differential symbolic execution.

Pasareanu, Corina S.↗
NASA NTRS1995

Fast Multipole Methods for Three-Dimensional N-body Problems

We are developing computational tools for the simulations of three-dimensional flows past bodies undergoing arbitrary motions. High resolution viscous vortex methods have been developed that allow for extended simulations of two-dimensional configurations such as vortex generators. Our objective is to extend this methodology to three dimensions and develop a robust computational scheme for the simulation of such flows. A fundamental issue in the use of vortex methods is the ability of employing efficiently large numbers of computational elements to resolve the large range of scales that exist in complex flows. The traditional cost of the method scales as Omicron (N(sup 2)) as the N computational elements/particles induce velocities at each other, making the method unacceptable for simulations involving more than a few tens of thousands of particles. In the last decade fast methods have been developed that have operation counts of Omicron (N log N) or Omicron (N) (referred to as BH and GR respectively) depending on the details of the algorithm. These methods are based on the observation that the effect of a cluster of particles at a certain distance may be approximated by a finite series expansion. In order to exploit this observation we need to decompose the element population spatially into clusters of particles and build a hierarchy of clusters (a tree data structure) - smaller neighboring clusters combine to form a cluster of the next size up in the hierarchy and so on. This hierarchy of clusters allows one to determine efficiently when the approximation is valid. This algorithm is an N-body solver that appears in many fields of engineering and science. Some examples of its diverse use are in astrophysics, molecular dynamics, micro-magnetics, boundary element simulations of electromagnetic problems, and computer animation. More recently these N-body solvers have been implemented and applied in simulations involving vortex methods. Koumoutsakos and Leonard (1995) implemented the GR scheme in two dimensions for vector computer architectures allowing for simulations of bluff body flows using millions of particles. Winckelmans presented three-dimensional, viscous simulations of interacting vortex rings, using vortons and an implementation of a BH scheme for parallel computer architectures. Bhatt presented a vortex filament method to perform inviscid vortex ring interactions, with an alternative implementation of a BH scheme for a Connection Machine parallel computer architecture.

Koumoutsakos, P.↗
DOE OSTI2024

Grain structure and texture selection regimes in metal powder bed fusion

Additive manufacturing (AM) offers opportunities to produce complex part geometries not possible with conventional processing and in some cases even improve part performance. However, adoption has been slowed by difficulties assessing microstructure variability and there is no straightforward approach to relate processing to grain structure characteristics. In this study, datasets from AdditiveFOAM heat transport simulations of laser powder bed fusion (LPBF) are used to drive ExaCA simulations of grain structure. The GPU utilization of ExaCA and an algorithmic update for modeling melt pool overlap region solidification enabled rapid and parallel simulation across a wider range of process conditions than previously explored with cellular automata-based solidification models. A texture selection angle $θ_s$ is defined based on melt pool overlap geometry, and the range of $θ_s$ over which a commonly observed texture transition occurs in characterized AM builds was well-reproduced by ExaCA simulations over a wide range of melt pool shape, hatch spacing, and layer height. ExaCA simulations with 90 degree rotation of the scan direction on every other layer reproduced a number of trends from the AM literature including grain refinement, the dominance of layers with larger melt pools on the final grain structure, and the weakening or strengthening of texture depending on odd and even layer melt pool overlap geometry. EBSD data from a benchmark AM part is used to validate the simulated mechanism of a layer rotation-induced texture strengthening effect. Importantly, these results expand the understanding of the mechanisms for texture selection in alloys with cubic crystal symmetry and offer an approach to easily evaluate processing conditions. With this new understanding, these modeling tools will enable anticipation of previously unexpected variations in grain structure and target specific microstructures and properties.

36 MATERIALS SCIENCE↗
DOE OSTI2018

Coping with silent and fail-stop errors at scale by combining replication and checkpointing

This paper provides a model and an analytical study of replication as a technique to cope with silent errors, as well as a mixture of both silent and fail-stop errors on large-scale platforms. Compared with fail-stop errors that are immediately detected when they occur, silent errors require a detection mechanism. To detect silent errors, many application-specific techniques are available, either based on algorithms (e.g., ABFT), invariant preservation or data analytics, but replication remains the most transparent and least intrusive technique. We explore the right level (duplication, triplication or more) of replication for two frameworks: (i) when the platform is subject to only silent errors, and (ii) when the platform is subject to both silent and fail-stop errors. A higher level of replication is more expensive in terms of resource usage but enables to tolerate more errors and to even correct some errors, hence there is a trade-off to be found. Replication is combined with checkpointing and comes with two flavors: process replication and group replication. Process replication applies to message-passing applications with communicating processes. Each process is replicated, and the platform is composed of process pairs, or triplets. Group replication applies to black-box applications, whose parallel execution is replicated several times. The platform is partitioned into two halves (or three thirds). In both scenarios, results are compared before each checkpoint, which is taken only when both results (duplication) or two out of three results (triplication) coincide. Otherwise, one or more silent errors have been detected, and the application rolls back to the last checkpoint, as well as when fail-stop errors have struck. We provide a detailed analytical study for all of these scenarios, with formulas to decide, for each scenario, the optimal parameters as a function of the error rate, checkpoint cost, and platform size. We also report a set of extensive simulation results that nicely corroborates the analytical model.

Checkpointing↗
NASA NTRS1991

Ionospheric refraction effects on TOPEX orbit determination accuracy using the Tracking and Data Relay Satellite System (TDRSS)

This investigation concerns the effects on Ocean Topography Experiment (TOPEX) spacecraft operational orbit determination of ionospheric refraction error affecting tracking measurements from the Tracking and Data Relay Satellite System (TDRSS). Although tracking error from this source is mitigated by the high frequencies (K-band) used for the space-to-ground links and by the high altitudes for the space-to-space links, these effects are of concern for the relatively high-altitude (1334 kilometers) TOPEX mission. This concern is due to the accuracy required for operational orbit-determination by the Goddard Space Flight Center (GSFC) and to the expectation that solar activity will still be relatively high at TOPEX launch in mid-1992. The ionospheric refraction error on S-band space-to-space links was calculated by a prototype observation-correction algorithm using the Bent model of ionosphere electron densities implemented in the context of the Goddard Trajectory Determination System (GTDS). Orbit determination error was evaluated by comparing parallel TOPEX orbit solutions, applying and omitting the correction, using the same simulated TDRSS tracking observations. The tracking scenarios simulated those planned for the observation phase of the TOPEX mission, with a preponderance of one-way return-link Doppler measurements. The results of the analysis showed most TOPEX operational accuracy requirements to be little affected by space-to-space ionospheric error. The determination of along-track velocity changes after ground-track adjustment maneuvers, however, is significantly affected when compared with the stringent 0.1-millimeter-per-second accuracy requirements, assuming uncoupled premaneuver and postmaneuver orbit determination. Space-to-space ionospheric refraction on the 24-hour postmaneuver arc alone causes 0.2 millimeter-per-second errors in along-track delta-v determination using uncoupled solutions. Coupling the premaneuver and postmaneuver solutions, however, appears likely to reduce this figure substantially. Plans and recommendations for response to these findings are presented.

Radomski, M. S.↗
NASA NTRS2013

ISSM: Ice Sheet System Model

In order to have the capability to use satellite data from its own missions to inform future sea-level rise projections, JPL needed a full-fledged ice-sheet/iceshelf flow model, capable of modeling the mass balance of Antarctica and Greenland into the near future. ISSM was developed with such a goal in mind, as a massively parallelized, multi-purpose finite-element framework dedicated to ice-sheet modeling. ISSM features unstructured meshes (Tria in 2D, and Penta in 3D) along with corresponding finite elements for both types of meshes. Each finite element can carry out diagnostic, prognostic, transient, thermal 3D, surface, and bed slope simulations. Anisotropic meshing enables adaptation of meshes to a certain metric, and the 2D Shelfy-Stream, 3D Blatter/Pattyn, and 3D Full-Stokes formulations capture the bulk of the ice-flow physics. These elements can be coupled together, based on the Arlequin method, so that on a large scale model such as Antarctica, each type of finite element is used in the most efficient manner. For each finite element referenced above, ISSM implements an adjoint. This adjoint can be used to carry out model inversions of unknown model parameters, typically ice rheology and basal drag at the ice/bedrock interface, using a metric such as the observed InSAR surface velocity. This data assimilation capability is crucial to allow spinning up of ice flow models using available satellite data. ISSM relies on the PETSc library for its vectors, matrices, and solvers. This allows ISSM to run efficiently on any parallel platform, whether shared or distrib- ISSM: Ice Sheet System Model NASA's Jet Propulsion Laboratory, Pasadena, California uted. It can run on the largest clusters, and is fully scalable. This allows ISSM to tackle models the size of continents. ISSM is embedded into MATLAB and Python, both open scientific platforms. This improves its outreach within the science community. It is entirely written in C/C++, which gives it flexibility in its design, and the power/speed that C/C++ allows. ISSM is svn (subversion) hosted, on a JPL repository, to facilitate its development and maintenance. ISSM can also model propagation of rifts using contact mechanics and mesh splitting, and can interface to the Dakota software. To carry out sensitivity analysis, mesh partitioning algorithms are available, based on the Scotch, Chaco, and Metis partitioners that ensure equal area mesh partitions can be done, which are then usable for sampling and local reliability methods.

Larour, Eric↗
DOE OSTI2026

Latent Twins

Over the past decade, scientific machine learning has transformed the development of mathematical and computational frameworks for analyzing, modeling, and predicting complex systems. From inverse problems to numerical partial differential equations (PDEs), dynamical systems, and model reduction, these advances have pushed the boundaries of what can be simulated. Yet they have often progressed in parallel, with representation learning and algorithmic solution methods evolving largely as separate pipelines. With Latent Twins, we propose a unifying mathematical framework that creates a hidden surrogate in latent space for the underlying equations. Whereas digital twins mirror physical systems in the digital world, Latent Twins mirror mathematical systems in a learned latent space governed by operators. Through this lens, classical modeling, inversion, model reduction, and operator approximation all emerge as special cases of a single principle. We establish the fundamental approximation properties of Latent Twins for both ordinary differential equations (ODEs) and PDEs and demonstrate the framework across three representative settings: (i) canonical ODEs, capturing diverse dynamical regimes; (ii) a PDE benchmark using the shallow-water equations, contrasting Latent Twin simulations with deep operator network and forecasts with a four-dimensional variational method baseline; and (iii) a challenging real-data geopotential reanalysis dataset, reconstructing and forecasting from sparse, noisy observations. Latent Twins provide a compact, interpretable surrogate for solution operators that evaluate across arbitrary time gaps in a single-shot, while remaining compatible with scientific pipelines such as assimilation, control, and uncertainty quantification. Looking forward, this framework offers scalable, theory-grounded surrogates that bridge data-driven representation learning and classical scientific modeling across disciplines.

Latent Twins↗
DOE OSTI2021

ArborX

ArborX library tackles a problem of efficiently finding geometric objects that are close in space. Variations of this problem, such as finding the nearest neighbors of a point, or finding all objects within a certain distance, are inherent components of applications in many fields. The data may be large so that solving the problem efficiently may require significant computational resources, such as multiple processors or accelerators such as general purpose GPUs. ArborX' main advantage in its ability to solve large problems efficiently utilizing a combination of distributed and on-node parallelism. ArborX can be run efficiently on a wide variety of hardware, including GPUs from different vendors, which distinguishes it from other available libraries which typically choose only few of these. The other advantage is that it supports both types of user problems: spatial problems (useful for intersections and finding objects within certain distance), and nearest neighbor problems. ArborX also supports flexible interface in its interaction with a user. Particularly, it allows a user to call user's own function on a positive match, a functionality not rarely available in other libraries. ArborX implements construction and traversal algorithms using efficient tree structures, such as bounding volume hierarchy (BVH). At its core, it uses linear BVH for its low construction cost and sufficient quality. ArborX is written using C++, and is parallelized using the message passing interface (MPI) for the distributed communication, and the Kokkos library for on-node parallelism. This approach allows ArborX to be run on a wide variety of hardware, from common laptops and desktops to supercomputers while using the same codebase. ArborX also implements several advanced algorithms using geometric search, such as density-based clustering algorithm DBSCAN.

ECP↗
DOE OSTI2025

ArborX 2.0

ArborX library tackles a problem of efficiently finding geometric objects that are close in space. Variations of this problem, such as finding the nearest neighbors of a point, or finding all objects within a certain distance, are inherent components of applications in many fields. The data may be large so that solving the problem efficiently may require significant computational resources, such as multiple processors or accelerators such as general purpose GPUs. ArborX' main advantage in its ability to solve large problems efficiently utilizing a combination of distributed and on-node parallelism. ArborX can be run efficiently on a wide variety of hardware, including GPUs from different vendors, which distinguishes it from other available libraries which typically choose only few of these. The other advantage is that it supports both types of user problems: spatial problems (useful for intersections and finding objects within certain distance), and nearest neighbor problems. ArborX also supports flexible interface in its interaction with a user. Particularly, it allows a user to call user's own function on a positive match, a functionality not rarely available in other libraries. ArborX implements construction and traversal algorithms using efficient tree structures, such as bounding volume hierarchy (BVH). At its core, ArborX uses linear BVH for its low construction cost and sufficient quality. ArborX implements both spatial and nearest-neighbor traversal algorithms. ArborX also provides several clustering algorithms (minimum spanning tree, DBSCAN, HDBSCAN*), interpolation using minimum least squares and ray tracing. ArborX is written using C++, and is parallelized using the message passing interface (MPI) for the distributed communication, and the Kokkos library for on-node parallelism. This approach allows ArborX to be run on a wide variety of hardware, from common laptops and desktops to supercomputers while using the same codebase.

Prokopenko, Andrey [Oak Ridge National Laboratory ↗
NASA NTRS1989

Multinode reconfigurable pipeline computer

A multinode parallel-processing computer is made up of a plurality of innerconnected, large capacity nodes each including a reconfigurable pipeline of functional units such as Integer Arithmetic Logic Processors, Floating Point Arithmetic Processors, Special Purpose Processors, etc. The reconfigurable pipeline of each node is connected to a multiplane memory by a Memory-ALU switch NETwork (MASNET). The reconfigurable pipeline includes three (3) basic substructures formed from functional units which have been found to be sufficient to perform the bulk of all calculations. The MASNET controls the flow of signals from the memory planes to the reconfigurable pipeline and vice versa. the nodes are connectable together by an internode data router (hyperspace router) so as to form a hypercube configuration. The capability of the nodes to conditionally configure the pipeline at each tick of the clock, without requiring a pipeline flush, permits many powerful algorithms to be implemented directly.

Nosenchuck, Daniel M.↗
DOE OSTI2023

Quantitative proton radiography and shadowgraphy for arbitrary intensities

Charged-particle radiography and shadowgraphy data can be directly inverted to obtain a line-integrated transverse Lorentz force or a line-integrated transverse refractive index gradient if intensity modulations due to scattering and absorption are negligible, and angular deflections are small. We develop a new direct-inversion algorithm based on plasma physics and compare it to a new Monge–Ampère code and an existing power diagram code. The measured or source intensity is represented by electrons subject to drag, and the other intensity by fixed ions. The decrease in kinetic plus electrostatic energy determines convergence. The displacement of the electrons from their initial to their equilibrium positions determines the line-integrated force or refractive index gradient. We have implemented two approaches: PIC (particle in cell) and Lagrangian fluid, in 1-D and 2-D. The PIC code works for arbitrary intensities, can work efficiently in parallel, and can make use of existing codes. The Lagrangian code requires less memory and is faster than the PIC code without massively parallel processing, but fails in 2-D for large intensity modulations. The Monge–Ampère code is by far the fastest in 2-D, without massively parallel processing, but fails for intensities with large voids, high contrast ratios and large deflections across the boundaries, and could not obtain the degree of convergence possible with the PIC code. As a result, the power diagram code was by far the slowest and most memory intensive, and failed for large peaks in the measured intensity.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2022

Bit-GraphBLAS: Bit-Level Optimizations of Matrix-Centric Graph Processing on GPU

In the graph data structure like adjacency matrix, the connectivity of two nodes can be sufficiently represented using only 1 bit, but they are generally treated as 32-bit full-precision in state-of-the-art graph frameworks to adopt common sparse format such as CSR. Meanwhile, bit-level parallelism has recently be explored to have high-performance potential and low storage requirement on GPUs with dense bit-tiles. To fill the gap, our solution is a hierarchical storage format that contains the bit-indexing base and dense bit-tile units. Inherently, the granularity of the bit-tile is an essential factor in achieving both storage compression and GPU parallelism. How to find a sweet spot that trades off between avoiding sparsity and exploiting is comprehensively researched in this work. In the experiment, we evaluate the proposed storage format and algorithms on modern generation GPUs, including Pascal and Volta, to figure out critical software co-designs in conjunction with existing hardware-specific optimization.

Chen, Jou-An↗