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 433 records · Page 24
All sourcesNASA NTRSDOE OSTI
NASA NTRS1989

Evaluating local indirect addressing in SIMD proc essors

In the design of parallel computers, there exists a tradeoff between the number and power of individual processors. The single instruction stream, multiple data stream (SIMD) model of parallel computers lies at one extreme of the resulting spectrum. The available hardware resources are devoted to creating the largest possible number of processors, and consequently each individual processor must use the fewest possible resources. Disagreement exists as to whether SIMD processors should be able to generate addresses individually into their local data memory, or all processors should access the same address. The tradeoff is examined between the increased capability and the reduced number of processors that occurs in this single instruction stream, multiple, locally addressed, data (SIMLAD) model. Factors are assembled that affect this design choice, and the SIMLAD model is compared with the bare SIMD and the MIMD models.

Middleton, David↗
NASA NTRSDate not supplied

Alternative Setup for Long-Duration Low-Duty-Cycle 600°C Ambient Testing of SiC Integrated Circuits

A scalable, compact oven testbed system for simultaneously evaluating a multitude of high temperature integrated circuits (ICs) for prolonged operating times of up to 600 °C has been prototyped. The new testbed system enables long-duration high temperature testing in sufficient statistical quantities consistent with standard aerospace electronics engineering standards. This setup is comprised of multiple compact ovens housing chips or packages mounted to ceramic circuit boards. Each oven is a compact 15.2 cm length by 15.2 cm width by 12.7 cm depth with a maximum 400 Watts of heating power. The custom-made silicon oxide ceramic heating block inside each oven is based on a 3D printed design adapted for the easy insertion of the IC device under test (DUT). This innovative design provides the quick insertion of ICs with or without a ceramic package into a 600 °C environment by utilizing a movable 11.43 cm long ceramic substrate with electrical traces extending from the oven hot zone to external standard plastic-based board connectors. Another key oven design feature is the minimization of the DUT exposure to electromagnetic interference (EMI) by utilizing a filtered DC power source to reduce heating element noise. Additionally, the ovens can be configured in a parallel arrangement allowing global data monitoring over a single industrial RS422 serial port. This feature is important for scaling up to test multiple ICs semi-simultaneously. A USB serial port is provided to independently control the operating parameters of each oven such as the oven target temperature and oven ramp rate. The oven temperature can reach up to 600 °C with a confirmed +/- 4 °C maximum deviation across the test zone region. The ramp rate can be programmed from 1 °C/minute up to 10 °C/minute. Furthermore, a programmable switchboard is used to interface with the DUT. This switchboard comprises a National Instruments Peripheral Component Interconnect eXtension for Instrumentation (NI PXI) system and a breakout board to send and receive power, analog or digital test signals. By using this unique oven testbed system, a variety of ICs can now be tested in parallel using the same test components configured for a diverse set of requirements.

DC Powered Oven↗
NASA NTRS1998

PLUM: Parallel Load Balancing for Unstructured Adaptive Meshes

Dynamic mesh adaption on unstructured grids is a powerful tool for computing large-scale problems that require grid modifications to efficiently resolve solution features. Unfortunately, an efficient parallel implementation is difficult to achieve, primarily due to the load imbalance created by the dynamically-changing nonuniform grid. To address this problem, we have developed PLUM, an automatic portable framework for performing adaptive large-scale numerical computations in a message-passing environment. First, we present an efficient parallel implementation of a tetrahedral mesh adaption scheme. Extremely promising parallel performance is achieved for various refinement and coarsening strategies on a realistic-sized domain. Next we describe PLUM, a novel method for dynamically balancing the processor workloads in adaptive grid computations. This research includes interfacing the parallel mesh adaption procedure based on actual flow solutions to a data remapping module, and incorporating an efficient parallel mesh repartitioner. A significant runtime improvement is achieved by observing that data movement for a refinement step should be performed after the edge-marking phase but before the actual subdivision. We also present optimal and heuristic remapping cost metrics that can accurately predict the total overhead for data redistribution. Several experiments are performed to verify the effectiveness of PLUM on sequences of dynamically adapted unstructured grids. Portability is demonstrated by presenting results on the two vastly different architectures of the SP2 and the Origin2OOO. Additionally, we evaluate the performance of five state-of-the-art partitioning algorithms that can be used within PLUM. It is shown that for certain classes of unsteady adaption, globally repartitioning the computational mesh produces higher quality results than diffusive repartitioning schemes. We also demonstrate that a coarse starting mesh produces high quality load balancing, at a fraction of the cost required a fine initial mesh. Results indicate that our parallel load balancing strategy will remain viable on large numbers of processors.

Oliker, Leonid↗
NASA NTRS1980

Parallel processing implementations of a contextual classifier for multispectral remote sensing data

Contextual classifiers are being developed as a method to exploit the spatial/spectral context of a pixel to achieve accurate classification. Classification algorithms such as the contextual classifier typically require large amounts of computation time. One way to reduce the execution time of these tasks is through the use of parallelism. The applicability of the CDC flexible processor system and of a proposed multimicroprocessor system (PASM) for implementing contextual classifiers is examined.

Siegel, H. J.↗
NASA NTRS2017

4BMS-X Design and Test Activation

In support of the NASA goals to reduce power, volume and mass requirements on future CO2 (Carbon Dioxide) removal systems for exploration missions, a 4BMS (Four Bed Molecular Sieve) test bed was fabricated and activated at the NASA Marshall Space Flight Center. The 4BMS-X (Four Bed Molecular Sieve-Exploration) test bed used components similar in size, spacing, and function to those on the flight ISS flight CDRA system, but were assembled in an open framework. This open framework allows for quick integration of changes to components, beds and material systems. The test stand is highly instrumented to provide data necessary to anchor predictive modeling efforts occurring in parallel to testing. System architecture and test data collected on the initial configurations will be presented.

Peters, Warren T.↗
NASA NTRS1994

Dynamic Load Balancing For Grid Partitioning on a SP-2 Multiprocessor: A Framework

Computational requirements of full scale computational fluid dynamics change as computation progresses on a parallel machine. The change in computational intensity causes workload imbalance of processors, which in turn requires a large amount of data movement at runtime. If parallel CFD is to be successful on a parallel or massively parallel machine, balancing of the runtime load is indispensable. Here a framework is presented for dynamic load balancing for CFD applications, called Jove. One processor is designated as a decision maker Jove while others are assigned to computational fluid dynamics. Processors running CFD send flags to Jove in a predetermined number of iterations to initiate load balancing. Jove starts working on load balancing while other processors continue working with the current data and load distribution. Jove goes through several steps to decide if the new data should be taken, including preliminary evaluate, partition, processor reassignment, cost evaluation, and decision. Jove running on a single IBM SP2 node has been completely implemented. Preliminary experimental results show that the Jove approach to dynamic load balancing can be effective for full scale grid partitioning on the target machine IBM SP2.

Sohn, Andrew↗
NASA NTRS1994

Dynamic Load Balancing for Grid Partitioning on a SP-2 Multiprocessor: A Framework

Computational requirements of full scale computational fluid dynamics change as computation progresses on a parallel machine. The change in computational intensity causes workload imbalance of processors, which in turn requires a large amount of data movement at runtime. If parallel CFD is to be successful on a parallel or massively parallel machine, balancing of the runtime load is indispensable. Here a framework is presented for dynamic load balancing for CFD applications, called Jove. One processor is designated as a decision maker Jove while others are assigned to computational fluid dynamics. Processors running CFD send flags to Jove in a predetermined number of iterations to initiate load balancing. Jove starts working on load balancing while other processors continue working with the current data and load distribution. Jove goes through several steps to decide if the new data should be taken, including preliminary evaluate, partition, processor reassignment, cost evaluation, and decision. Jove running on a single EBM SP2 node has been completely implemented. Preliminary experimental results show that the Jove approach to dynamic load balancing can be effective for full scale grid partitioning on the target machine IBM SP2.

Sohn, Andrew↗
NASA NTRS1994

Dynamic Load Balancing for Finite Element Calculations on Parallel Computers

Computational requirements of full scale computational fluid dynamics change as computation progresses on a parallel machine. The change in computational intensity causes workload imbalance of processors, which in turn requires a large amount of data movement at runtime. If parallel CFD is to be successful on a parallel or massively parallel machine, balancing of the runtime load is indispensable. Here a frame work is presented for dynamic load balancing for CFD applications, called Jove. One processor is designated as a decision maker Jove while others are assigned to computational fluid dynamics. Processors running CFD send flags to Jove in a predetermined number of iterations to initiate load balancing. Jove starts working on load balancing while other processors continue working with the current data and load distribution. Jove goes through several steps to decide if the new data should be taken, including preliminary evaluate, partition, processor reassignment, cost evaluation, and decision. Jove running on a single SP2 node has been completely implemented. Preliminary experimental results show that the Jove approach to dynamic load balancing can be effective for full scale grid partitioning on the target machine SP2.

Pramono, Eddy↗
NASA NTRS1993

Feed-forward volume rendering algorithm for moderately parallel MIMD machines

Algorithms for direct volume rendering on parallel and vector processors are investigated. Volumes are transformed efficiently on parallel processors by dividing the data into slices and beams of voxels. Equal sized sets of slices along one axis are distributed to processors. Parallelism is achieved at two levels. Because each slice can be transformed independently of others, processors transform their assigned slices with no communication, thus providing maximum possible parallelism at the first level. Within each slice, consecutive beams are incrementally transformed using coherency in the transformation computation. Also, coherency across slices can be exploited to further enhance performance. This coherency yields the second level of parallelism through the use of the vector processing or pipelining. Other ongoing efforts include investigations into image reconstruction techniques, load balancing strategies, and improving performance.

Yagel, Roni↗
NASA NTRS1984

Plasma and energetic particle structure upstream of a quasi-parallel interplanetary shock

ISEE 1, 2 and 3 data from 1978 on interplanetary magnetic fields, shock waves and particle energetics are examined to characterize a quasi-parallel shock. The intense shock studied exhibited a 640 km/sec velocity. The data covered 1-147 keV protons and electrons and ions with energies exceeding 30 keV in regions both upstream and downstream of the shock, and also the magnitudes of ion-acoustic and MHD waves. The energetic particles and MHD waves began being detected 5 hr before the shock. Intense halo electron fluxes appeared ahead of the shock. A closed magnetic field structure was produced with a front end 700 earth radii from the shock. The energetic protons were cut off from the interior of the magnetic bubble, which contained a markedly increased density of 2-6 keV protons as well as the shock itself.

Kennel, C. F.↗
NASA NTRS1984

Data flow modeling techniques

There have been a number of simulation packages developed for the purpose of designing, testing and validating computer systems, digital systems and software systems. Complex analytical tools based on Markov and semi-Markov processes have been designed to estimate the reliability and performance of simulated systems. Petri nets have received wide acceptance for modeling complex and highly parallel computers. In this research data flow models for computer systems are investigated. Data flow models can be used to simulate both software and hardware in a uniform manner. Data flow simulation techniques provide the computer systems designer with a CAD environment which enables highly parallel complex systems to be defined, evaluated at all levels and finally implemented in either hardware or software. Inherent in data flow concept is the hierarchical handling of complex systems. In this paper we will describe how data flow can be used to model computer system.

Kavi, K. M.↗
NASA NTRSDate not supplied

MLtool++ package for machine learning and its applications to materials data

We are developing Mltool++ package of software programs for machine learning (ML). Given the MLtool Python code, we create a faster C++ code with the potential for parallelization. We have extracted materials data from the literature. One dataset contains melting temperatures of stoichiometric 1:1 metallic compounds XZ, composed by elements X={Al, Ti, V, Cr, Zr, Nb, Mo, Hf, Ta, W} and Z={Co, Ni, Cu, Rh, Pd, Ag, Ir, Pt, Au}, and another contains solid-solid symmetry-breaking phase transition temperatures. We studied dependences of temperatures on composition, found several correlations, and parametrized them by analytical functions. Mltool++ package is generic and applicable to any tabulated numeric data.

Pierce M. Pettit↗
NASA NTRS1965

Logic redundancy improves digital system reliability

Redundant-channel system automatically corrects any single error in a set of three binary signal channels. This system is especially applicable to digital computers where data is transmitted in parallel channels.

Source record↗
NASA NTRS1975

Investigation of system integration methods for bubble domain flight recorders

System integration methods for bubble domain flight records are investigated. Bubble memory module packaging and assembly, the control electronics design and construction, field coils, and permanent magnet bias structure design are studied. A small 60-k bit engineering model was built and tested to demonstrate the feasibility of the bubble recorder. Based on the various studies performed, a projection is made on a 50,000,000-bit prototype recorder. It is estimated that the recorder will occupy 190 cubic in., weigh 12 lb, and consume 12 w power when all of its four tracks are operated in parallel at 150 kHz data rate.

Chen, T. T.↗
NASA NTRS1983

Parallel processing in a host plus multiple array processor system for radar

Host plus multiple array processor architecture is demonstrated to yield a modular, fast, and cost-effective system for radar processing. Software methodology for programming such a system is developed. Parallel processing with pipelined data flow among the host, array processors, and discs is implemented. Theoretical analysis of performance is made and experimentally verified. The broad class of problems to which the architecture and methodology can be applied is indicated.

Barkan, B. Z.↗
NASA NTRS1983

Cumulative reports and publications through 31 December 1983

All reports for the calendar years 1975 through December 1983 are listed by author. Since ICASE reports are intended to be preprints of articles for journals and conference proceedings, the published reference is included when available. Thirteen older journal and conference proceedings references are included as well as five additional reports by ICASE personnel. Major categories of research covered include: (1) numerical methods, with particular emphasis on the development and analysis of basic algorithms; (2) computational problems in engineering and the physical sciences, particularly fluid dynamics, acoustics, structural analysis, and chemistry; and (3) computer systems and software, especially vector and parallel computers, microcomputers, and data management.

Source record↗
NASA NTRS1987

Parallel Analog-to-Digital Image Processor

Proposed integrated-circuit network of many identical units convert analog outputs of imaging arrays of x-ray or infrared detectors to digital outputs. Converter located near imaging detectors, within cryogenic detector package. Because converter output digital, lends itself well to multiplexing and to postprocessing for correction of gain and offset errors peculiar to each picture element and its sampling and conversion circuits. Analog-to-digital image processor is massively parallel system for processing data from array of photodetectors. System built as compact integrated circuit located near local plane. Buffer amplifier for each picture element has different offset.

Lokerson, D. C.↗
NASA NTRS1987

Effect of an oxygen plasma on uncoated thin aluminum reflecting films

Thin aluminum films were considered for use as a reflective surface for solar collectors on orbiting solar dynamic power systems. A matter of concern is the durability of such reflective coatings against oxidative attack by highly reactive neutral atomic oxygen, which is the predominate chemical specie in low Earth orbit. Research to date was aimed at evaluating the protective merit of thin dielectric coatings over the aluminum or other reflective metals. However, an uncoated aluminum reflector may self-protect by virtue of the oxide formed from its exposed surface, which constitutes a physical barrier to further oxidation. This possibility was investigated, and an attempt was made to characterize the effects of atomic oxygen on thin Al films using photomicrographs, scanning electron microscopy, spectrophotometry, Auger analysis, and mass measurements. Data collected in a parallel effort is discussed for its comparative value. The results of the investigation of uncoated aluminum supported the self-protection hypothesis, and importantly, it was found that long term specular reflectance for uncoated aluminum exceeded that of Al and Ag reflectors with dielectric coatings.

Parsons, Roger L.↗