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 1,657 records · Page 92

Pilot study for compact microbeam radiation therapy using a carbon nanotube field emission micro-CT scanner

Purpose: Microbeam radiation therapy (MRT) is defined as the use of parallel, microplanar x-ray beams with an energy spectrum between 50 and 300 keV for cancer treatment and brain radiosurgery. Up until now, the possibilities of MRT have mainly been studied using synchrotron sources due to their high flux (100s Gy/s) and approximately parallel x-ray paths. The authors have proposed a compact x-ray based MRT system capable of delivering MRT dose distributions at a high dose rate. This system would employ carbon nanotube (CNT) field emission technology to create an x-ray source array that surrounds the target of irradiation. Using such a geometry, multiple collimators would shape the irradiation from this array into multiple microbeams that would then overlap or interlace in the target region. This pilot study demonstrates the feasibility of attaining a high dose rate and parallel microbeam beams using such a system. Methods: The microbeam dose distribution was generated by our CNT micro-CT scanner (100μm focal spot) and a custom-made microbeam collimator. An alignment assembly was fabricated and attached to the scanner in order to collimate and superimpose beams coming from different gantry positions. The MRT dose distribution was measured using two orthogonal radiochromic films embedded inside a cylindrical phantom. This target was irradiated with microbeams incident from 44 different gantry angles to simulate an array of x-ray sources as in the proposed compact CNT-based MRT system. Finally, phantom translation in a direction perpendicular to the microplanar beams was used to simulate the use of multiple parallel microbeams. Results: Microbeams delivered from 44 gantry angles were superimposed to form a single microbeam dose distribution in the phantom with a FWHM of 300μm (calculated value was 290 μm). Also, during the multiple beam simulation, a peak to valley dose ratio of ∼10 was found when the phantom translation distance was roughly 4x the beam width. The first prototype CNT-based x-ray tube dedicated to the development of compact MRT technology development was proposed and planned based on the preliminary experimental results presented here and the previous corresponding Monte Carlo simulations. Conclusions: The authors have demonstrated the feasibility of creating microbeam dose distributions at a high dose rate using a proposed compact MRT system. The flexibility of CNT field emission x-ray sources could possibly bring compact and low cost MRT devices to the larger research community and assist in the translational research of this promising new approach to radiation therapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

A low-communication, parallel algorithm for solving PDEs based on range decomposition

This work proposes a new, low-communication algorithm for solving PDEs on massively parallel computers. The range decomposition (RD) algorithm exposes coarse-grain parallelism by applying nested iteration and adaptive mesh refinement locally before performing a global communication step. Just a few such steps are observed to be sufficient to obtain accuracy within a small multiple of discretization error. The target applications are petascale and exascale machines, where hierarchical parallelism is required and traditional parallel numerical PDE communication patterns are costly because of message latency. The RD algorithm uses a partition of unity to equally distribute the error, and thus, the work. The computational advantages of this approach are that the decomposed problems can be solved in parallel without any communication until the partitioned solutions are summed. This offers potential advantages in the paradigm of expensive communication but very cheap computation. This paper introduces the method and explains the details of the communication step. Two performance models are developed, showing that the latency cost associated with a traditional parallel implementation of nested iteration is proportional to log(P) 2 , whereas the RD method reduces the communication latency to log(P), while maintaining similar bandwidth costs. Numerical results for two problems, Laplace and advection diffusion, demonstrate the enhanced performance, and a heuristic argument explains why the method converges quickly.

97 MATHEMATICS AND COMPUTING↗

Ulysses field and plasma observations of magnetic holes in the solar wind and their relation to mirror-mode structures

The term 'magnetic hole' has been used to denote isolated intervals when the magnitude of the interplanetary magnetic field drops to a few tenths, or less, of its ambient value for a time that corresponds to a linear dimension of tens to a few hundreds of proton gyro-radii. Data obtained by the Ulysses magnetometer and solar wind anlayzer have been combined to study the properties of such magnetic holes in the solar wind between 1 AU and 5.4 AU and to 23 deg south latitude. In order to avoid confusion with decreases in field strength at interplanetary discontinuities, the study has focused on linear holes across which the field direction changed by less than 5 deg. The holes occurred preferentially, but not without exception, in the interaction regions on the leading edges of high-speed solar wind streams. Although the plasma surrounding the holes was generally stable against the mirror instability, there are indications that the holes may have been remnants of mirror-mode structures created upstream of the points of observation. Those indications include the following: (1) For the few holes for which proton of alpha-particle pressure could be measured inside the hole, the ion thermal pressure was always greater than in the plasma adjacent to the holes. (2) The plasma surrounding many of the holes was marginally stable for the mirror mode, while the plasma environment of all holes was significantly closer to mirror instability than was the average solar wind. (3) The plasma containing trains of closely spaced holes was closer to mirror instability than was the plasma containing isolated holes. (4) The near-hole plasma had much higher ion beta (ratio of thermal to magnetic pressure) than did the average solar wind. (5) Near the holes, T(sub perp)/T(sub parallel) tended to be either greater than 1 or larger than in the average wind. (6) The proton and alpha-particle distribution functions measured inside the holes occasionally exhibited the flattened phase-space-density contoures in nu(sub perp)/nu(sub parallel) space found in some numerical simulations of the mirror instability.

Winterhalter, Daniel↗

Distributed Outage Detection in Power Distribution Networks

Real time topology knowledge is essential for situational awareness of power distribution networks. Line outages change the topology of a distribution network. Hence, outage detection is an important task. Most of the existing outage detection algorithms are centralized, in which sensors communicate their data to a control center which performs outage detection using the received data. However, with the increasing size of the distribution network and with different areas of the network being monitored by different operators, communication is a bottleneck and scalability is a major concern. To address these issues, we propose a novel outage detection algorithm using a divide and conquer approach. First, we divide a distribution network into sub-networks, such that outage detection can be run in parallel in each sub-network independently ensuring scalability to large networks. Further, to reduce the latency, bandwidth and attenuation challenges associated with communications in a large network, we divide each sub-network into multiple control areas which communicate only with their neighbors. We employ a distributed iterative load estimation across the control areas of each sub-network and then use the load estimate for local outage detection in each control area. Here, the performance of our algorithm is evaluated for multiple feeder models and compared against traditional centralized outage detection algorithms.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Design and Implementation of a Distributed Version of the NASA Engine Performance Program

Distributed NEPP is a new version of the NASA Engine Performance Program that runs in parallel on a collection of Unix workstations connected through a network. The program is fault-tolerant, efficient, and shows significant speed-up in a multi-user, heterogeneous environment. This report describes the issues involved in designing distributed NEPP, the algorithms the program uses, and the performance distributed NEPP achieves. It develops an analytical model to predict and measure the performance of the simple distribution, multiple distribution, and fault-tolerant distribution algorithms that distributed NEPP incorporates. Finally, the appendices explain how to use distributed NEPP and document the organization of the program's source code.

Cours, Jeffrey T.↗

Electromagnetic Design of a Magnetically-Coupled Spatial Power Combiner

The design of a two-dimensional beam-combining network employing a parallel-plate superconducting waveguide with a mono-crystalline silicon dielectric is presented. This novel beam-combining network structure employs an array of magnetically coupled antenna elements to achieve high coupling efficiency and full sampling of the intensity distribution while avoiding diffractive losses in the multi-mode region defined by the parallel-plate waveguide. These attributes enable the structures use in realizing compact far-infrared spectrometers for astrophysical and instrumentation applications. When configured with a suitable corporate-feed power-combiner, this fully sampled array can be used to realize a low-sidelobe apodized response without incurring a reduction in coupling efficiency. To control undesired reflections over a wide range of angles in the finite-sized parallel-plate waveguide region, a wideband meta-material electromagnetic absorber structure is implemented. This adiabatic structure absorbs greater than 99 of the power over the 1.7:1 operational band at angles ranging from normal (0 degree) to near parallel (180 degree) incidence. Design, simulations, and application of the device will be presented.

Bulcha, B.↗

Dynamic Load Balancing for Adaptive Computations on Distributed-Memory Machines

Dynamic load balancing is central to adaptive mesh-based computations on large-scale parallel computers. The principal investigator has investigated various issues on the dynamic load balancing problem under NASA JOVE and JAG rants. The major accomplishments of the project are two graph partitioning algorithms and a load balancing framework. The S-HARP dynamic graph partitioner is known to be the fastest among the known dynamic graph partitioners to date. It can partition a graph of over 100,000 vertices in 0.25 seconds on a 64- processor Cray T3E distributed-memory multiprocessor while maintaining the scalability of over 16-fold speedup. Other known and widely used dynamic graph partitioners take over a second or two while giving low scalability of a few fold speedup on 64 processors. These results have been published in journals and peer-reviewed flagship conferences.

Source record↗

Laser-induced nonsequential double ionization at and above the recollision-excitation-tunneling threshold

We perform a rigorous, semianalytic study of the recollision excitation with subsequent tunneling ionization (RESI) mechanism in laser-induced nonsequential double ionization (NSDI), based on the strong-field approximation. We show that the shapes of the electron momentum distributions carry information about the bound state with which the first electron collides, the bound state to which the second electron is excited, and the type of electron-electron interaction. Furthermore, one may define a driving-field intensity threshold for the RESI physical mechanism. At the threshold, the kinetic energy of the first electron, upon return, is just sufficient to excite the second electron. We compute the distributions for helium and argon in the threshold and above-threshold intensity regimes. In the latter case, we relate our findings to existing experiments. The electron momentum distributions encountered are symmetric with respect to all quadrants of the plane spanned by the momentum components parallel to the laser-field polarization, instead of concentrating on only the second and fourth quadrants. The above-mentioned momentum constraints, together with the strong dependence of the distributions on the bound states involved, may be important for singling out the RESI mechanism in actual physical situations and using NSDI in ultrafast imaging.

74 ATOMIC AND MOLECULAR PHYSICS↗

Excitation of a plasmon resonance in metal cylinders by an evanescent wave

The magnetic field distribution and electric-field amplification factor are found in the gap between periodically arranged parallel metal cylinders scattering an evanescent wave. Such a wave appears if an original plane wave is incident from the dielectric substrate onto the interface at an angle of total internal reflection. A substantial restructuring of the near-field distribution, which results from a small change in the angle of incidence, is revealed. Estimations of the angular dependence of the local field, explaining the shift and broadening of the plasmon resonance at a shorter wavelength, are presented. (nanostructures)

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Power conditioning for large dc motors for space flight applications

The design and performance of a prototype power-conditioning system for use with large brushless dc motors on NASA space missions are discussed in detail and illustrated with extensive diagrams, drawings, and graphs. The 5-kW 8-phase parallel module evaluated here would be suitable for use in the Space Shuttle Orbiter cargo bay. A current-balancing magnetic assembly with low distributed inductance permits high-speed current switching from a low-voltage bus as well as current balancing between parallel MOSFETs.

Veatch, Martin S.↗

Electron energy distributions measured during electron beam/plasma interactions

In the large vacuum facility at the NASA-Johnson Space Center an electron beam was projected 20 m parallel to B from a gun with variable accelerating potential (1.0 to 2.5 kV) to an aluminum target. The ionospheric neutral pressure and field were approximated. Beam electron energy distributions were measured directly using an electrostatic deflection analyzer and indirectly with a detector that responded to the X-rays produced by electron impact on the target. At low currents the distribution is sharply peaked at the acceleration potential. At high currents a beam plasma discharge occurs and electrons are redistributed in energy so that the former energy peak broadens to 10-15 percent FWHM with a strongly enhanced low energy tail. At the 10% of maximum point the energy spectrum ranges from less than 1/2 to 1.2 times the gun energy. The effect is qualitatively the same at all pitch angles and locations sampled.

Jost, R. J.↗

Proposal for massively parallel data storage system

An architecture for integrating large numbers of data storage units (drives) to form a distributed mass storage system is proposed. The network of interconnected units consists of nodes and links. At each node there resides a controller board, a data storage unit and, possibly, a local/remote user-terminal. The links (twisted-pair wires, coax cables, or fiber-optic channels) provide the communications backbone of the network. There is no central controller for the system as a whole; all decisions regarding allocation of resources, routing of messages and data-blocks, creation and distribution of redundant data-blocks throughout the system (for protection against possible failures), frequency of backup operations, etc., are made locally at individual nodes. The system can handle as many user-terminals as there are nodes in the network. Various users compete for resources by sending their requests to the local controller-board and receiving allocations of time and storage space. In principle, each user can have access to the entire system, and all drives can be running in parallel to service the requests for one or more users. The system is expandable up to a maximum number of nodes, determined by the number of routing-buffers built into the controller boards. Additional drives, controller-boards, user-terminals, and links can be simply plugged into an existing system in order to expand its capacity.

Mansuripur, M.↗

Debugging tasked Ada programs

The applications for which Ada was developed require distributed implementations of the language and extensive use of tasking facilities. Debugging and testing technology as it applies to parallel features of languages currently falls short of needs. Thus, the development of embedded systems using Ada pose special challenges to the software engineer. Techniques for distributing Ada programs, support for simulating distributed target machines, testing facilities for tasked programs, and debugging support applicable to simulated and to real targets all need to be addressed. A technique is presented for debugging Ada programs that use tasking and it describes a debugger, called AdaTAD, to support the technique. The debugging technique is presented together with the use interface to AdaTAD. The component of AdaTAD that monitors and controls communication among tasks was designed in Ada and is presented through an example with a simple tasked program.

Fainter, R. G.↗

Computational methods for analyzing the transmission characteristics of a beta particle magnetic analysis system

Computational methods were developed to study the trajectories of beta particles (positrons) through a magnetic analysis system as a function of the spatial distribution of the radionuclides in the beta source, size and shape of the source collimator, and the strength of the analyzer magnetic field. On the basis of these methods, the particle flux, their energy spectrum, and source-to-target transit times have been calculated for Na-22 positrons as a function of the analyzer magnetic field and the size and location of the target. These data are in studies requiring parallel beams of positrons of uniform energy such as measurement of the moisture distribution in composite materials. Computer programs for obtaining various trajectories are included.

Singh, J. J.↗

SELF-CONSISTENT ION CYCLOTRON ANISOTROPY-BETA RELATION FOR SOLAR WIND PROTONS

We derive a set of self-consistent marginally stable states for a system of ion-cyclotron waves propagating parallel to the large-scale magnetic field through a homogeneous proton-electron plasma. The proton distributions and the wave dispersions are related through the condition that no further ion-cyclotron resonant particle scattering or wave growth/damping may take place. The thermal anisotropy of the protons in these states therefore defines the threshold value for triggering the proton-cyclotron anisotropy instability. A number of recent papers have noted that the anisotropy of solar wind protons at 1 AU does not seem to be limited by the proton-cyclotron anisotropy threshold, even at low plasma beta. However, this puzzle seems to be due solely to the estimation of this anisotropy threshold under the assumption that the protons have a bi-Maxwellian distribution. We note that bi-Maxwellian distributions are never marginally stable to the resonant cyclotron interaction, so these estimates do not represent physically valid thresholds. The threshold anisotropies obtained from our marginally stable states are much larger, as a function of proton parallel beta, than the bi-Maxwellian estimates, and we show that the measured data remains below these more rigorous thresholds. Thus, the results of this paper resolve the apparent contradiction presented by the solar wind anisotropy observations at 1 AU: the bi-Maxwellian anisotropies are not rigorous thresholds, and so do not limit the proton distributions in the solar wind.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Paired organic matter and pyrite δ 34 S records reveal mechanisms of carbon, sulfur, and iron cycle disruption during Ocean Anoxic Event 2

The sulfur (S) isotope composition of pyrite in the sedimentary record has played an important part in our understanding of the evolution of biogeochemical cycles throughout Earth history. However, the kinetics of pyritization are complex and depend strongly on the reactivity and mineralogy of available iron. As a second major sink for sulfide in anoxic sediments, organic matter (OM) provides essential context for reconstructing the distribution and isotopic composition of environmental sulfide. To first order, roughly parallel pyrite and OM δ 34 S profiles reflect changes in sulfide, while independent patterns require alternative explanations, including changes in iron availability or OM characteristics. We apply this framework to Ocean Anoxic Event 2 (OAE-2, ~94 Mya), a period of enhanced burial of reduced C and S (in OM and pyrite) that has been associated with an expansion of reducing marine conditions. We present paired S-isotope records for pyrite and OM along with profiles of OM S:C ratio and S redox speciation from four well-characterized lithologic sections with a range of depositional environments (Pont d'Issole, Cismon, Tarfaya Basin, and Demerara Rise) to reconstruct both local redox structure and global mechanisms impacting the C, S and Fe cycles around OAE-2.

54 ENVIRONMENTAL SCIENCES↗

Computing the ankle-brachial index with parallel computational fluid dynamics

The ankle-brachial index (ABI), a ratio of arterial blood pressure in the ankles and upper arms, is used to diagnose and monitor circulatory conditions such as coarctation of the aorta and peripheral artery disease. Computational simulations of the ABI can potentially determine the parameters that produce an ABI indicative of ischemia or other abnormalities in blood flow. However, 0- and 1-D computational methods are limited in describing a 3-D patient-derived geometry. Thus, we present a massively parallel framework for computational fluid dynamics (CFD) simulations in the full arterial system. Additionally, using the lattice Boltzmann method to solve the Navier–Stokes equations, we employ highly parallelized and scalable methods to generate the simulation domain and efficiently distribute the computational load among processors. For the first time, we compute an ABI with 3-D CFD. In this proof-of-concept study, we investigate the dependence of ABI on the presence of stenoses, or narrowed regions of the arteries, by directly modifying the arterial geometry. As a result, our framework enables the computation a hemodynamic factor characterizing flow at the scale of the full arterial system, in a manner that is extensible to patient-specific imaging data and holds potential for treatment planning.

59 BASIC BIOLOGICAL SCIENCES↗