Search NASA⌕ Search

SEARCH · Search NASA

Results for “Computational Load Balancing”

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 145 records · Page 8

A simple method to compute the change in earth-atmosphere radiative balance due to a stratospheric aerosol layer

A computer code was developed in terms of a three-layer model for the earth-atmosphere system, using a two-stream approximation for the troposphere and stratosphere. The analysis was limited to variable atmosphere loading by solar radiation over an unperturbed section of the atmosphere. The scattering atmosphere above a Lambertian ground layer was considered in order to derive the planar albedo and the spherical albedo. Attention was given to the influence of the aerosol optical thickness in the stratosphere, the single scattering albedo and asymmetry factor, and the sublayer albedo. Calculations were performed of the zonal albedo and the planetary radiation balance, taking into account a stratospheric aerosol layer containing H2SO4 droplets and volcanic ash. The resulting ground temperature disturbance was computed using a Budyko (1969) climate model. Local decreases in the albedo in the summer were observed in high latitudes, implying a heating effect of the aerosol. An accompanying energy loss of 23-27 W/sq m was projected, which translates to surface temperature decreases of either 1.1 and 0.45 C, respectively, for background and volcanic aerosols.

Lenoble, J.↗

Cartesian Off-Body Grid Adaption for Viscous Time- Accurate Flow Simulation

An improved solution adaption capability has been implemented in the OVERFLOW overset grid CFD code. Building on the Cartesian off-body approach inherent in OVERFLOW and the original adaptive refinement method developed by Meakin, the new scheme provides for automated creation of multiple levels of finer Cartesian grids. Refinement can be based on the undivided second-difference of the flow solution variables, or on a specific flow quantity such as vorticity. Coupled with load-balancing and an inmemory solution interpolation procedure, the adaption process provides very good performance for time-accurate simulations on parallel compute platforms. A method of using refined, thin body-fitted grids combined with adaption in the off-body grids is presented, which maximizes the part of the domain subject to adaption. Two- and three-dimensional examples are used to illustrate the effectiveness and performance of the adaption scheme.

Buning, Pieter G.↗

Collectives for Multiple Resource Job Scheduling Across Heterogeneous Servers

Efficient management of large-scale, distributed data storage and processing systems is a major challenge for many computational applications. Many of these systems are characterized by multi-resource tasks processed across a heterogeneous network. Conventional approaches, such as load balancing, work well for centralized, single resource problems, but breakdown in the more general case. In addition, most approaches are often based on heuristics which do not directly attempt to optimize the world utility. In this paper, we propose an agent based control system using the theory of collectives. We configure the servers of our network with agents who make local job scheduling decisions. These decisions are based on local goals which are constructed to be aligned with the objective of optimizing the overall efficiency of the system. We demonstrate that multi-agent systems in which all the agents attempt to optimize the same global utility function (team game) only marginally outperform conventional load balancing. On the other hand, agents configured using collectives outperform both team games and load balancing (by up to four times for the latter), despite their distributed nature and their limited access to information.

Tumer, K.↗

Diffusion freezeout in gas-loaded heat pipes.

Experimental diffusion freezeout rates in a gas loaded heat pipe are presented and compared with predictions of an analysis and computer program published earlier. The experimental approach was to pivot a heat pipe on knife edges and use an analytical mass balance to measure the rate of mass migration within the pipe. Diffusion freezeout rates in a water heat pipe were measured as a function of the condenser sink temperature and the condenser-to-sink thermal conductance. Measured rates varied between 19 and 270 mg/hr and yielded 64 to 99 percent agreement with predicted values. These results indicated that the analysis and computer program for predicting the behavior of gas loaded heat pipes are useful tools for designing such systems.

Marcus, B. D.↗

A Portable MPI Implementation of the SPAI Preconditioner in ISIS++

A parallel MPI implementation of the Sparse Approximate Inverse (SPAI) preconditioner is described. SPAI has proven to be a highly effective preconditioner, and is inherently parallel because it computes columns (or rows) of the preconditioning matrix independently. However, there are several problems that must be addressed for an efficient MPI implementation: load balance, latency hiding, and the need for one-sided communication. The effectiveness, efficiency, and scaling behavior of our implementation will be shown for different platforms.

Barnard, Stephen T.↗

Applications of concurrent neuromorphic algorithms for autonomous robots

This article provides an overview of studies at the Oak Ridge National Laboratory (ORNL) of neural networks running on parallel machines applied to the problems of autonomous robotics. The first section provides the motivation for our work in autonomous robotics and introduces the computational hardware in use. Section 2 presents two theorems concerning the storage capacity and stability of neural networks. Section 3 presents a novel load-balancing algorithm implemented with a neural network. Section 4 introduces the robotics test bed now in place. Section 5 concerns navigation issues in the test-bed system. Finally, Section 6 presents a frequency-coded network model and shows how Darwinian techniques are applied to issues of parameter optimization and on-line design.

Barhen, J.↗

Comparing the Performance of Two Dynamic Load Distribution Methods

Parallel processing of symbolic computations on a message-passing multi-processor presents one challenge: To effectively utilize the available processors, the load must be distributed uniformly to all the processors. However, the structure of these computations cannot be predicted in advance. go, static scheduling methods are not applicable. In this paper, we compare the performance of two dynamic, distributed load balancing methods with extensive simulation studies. The two schemes are: the Contracting Within a Neighborhood (CWN) scheme proposed by us, and the Gradient Model proposed by Lin and Keller. We conclude that although simpler, the CWN is significantly more effective at distributing the work than the Gradient model.

Kale, L. V.↗

Investigation of Unsteady Pressure-Sensitive Paint (uPSP) and a Dynamic Loads Balance to Predict Launch Vehicle Buffet Environments

This NESC assessment examined the accuracy of estimating buffet loads on in-line launch vehicles without booster attachments using sparse unsteady pressure measurements. The buffet loads computed using sparse sensor data were compared with estimates derived using measurements with much higher spatial resolution. The current method for estimating launch vehicle buffet loads is through wind tunnel testing of models with approximately 400 unsteady pressure transducers. Even with this relatively large number of sensors, the coverage can be insufficient to provide reliable integrated unsteady loads on vehicles. In general, sparse sensor spacing requires the use of coherence-length-based corrections in the azimuthal and axial directions to integrate the unsteady pressures and obtain reasonable estimates of the buffet loads. Coherence corrections have been used to estimate buffet loads for a variety of launch vehicles with the assumption methodology results in reasonably conservative loads. For the Space Launch System (SLS), the first estimates of buffet loads exceeded the limits of the vehicle structure, so additional tests with higher sensor density were conducted to better define the buffet loads and possibly avoid expensive modifications to the vehicle design. Without the additional tests and improvements to the coherence-length analysis methods, there would have been significant impacts to the vehicle weight, cost, and schedule. If the load estimates turn out to be too low, there is significant risk of structural failure of the vehicle. This assessment used a combination of unsteady pressure-sensitive paint (uPSP), unsteady pressure transducers, and a dynamic force and moment balance to investigate the integration schemes used with limited unsteady pressure data by comparing them with direct integration of extremely dense fluctuating pressure measurements. An outfall of the assessment was to evaluate the potential of using the emerging uPSP technique in a production test environment for future launch vehicles. The results show that modifications to the current technique can improve the accuracy of buffet estimates. More importantly, the uPSP worked remarkably well and, with improvements to the frequency response, sensitivity, and productivity, will provide an enhanced method for measuring wind tunnel buffet forcing functions (BFFs).

Schuster, David M.↗

A nonrecursive order N preconditioned conjugate gradient: Range space formulation of MDOF dynamics

While excellent progress has been made in deriving algorithms that are efficient for certain combinations of system topologies and concurrent multiprocessing hardware, several issues must be resolved to incorporate transient simulation in the control design process for large space structures. Specifically, strategies must be developed that are applicable to systems with numerous degrees of freedom. In addition, the algorithms must have a growth potential in that they must also be amenable to implementation on forthcoming parallel system architectures. For mechanical system simulation, this fact implies that algorithms are required that induce parallelism on a fine scale, suitable for the emerging class of highly parallel processors; and transient simulation methods must be automatically load balancing for a wider collection of system topologies and hardware configurations. These problems are addressed by employing a combination range space/preconditioned conjugate gradient formulation of multi-degree-of-freedom dynamics. The method described has several advantages. In a sequential computing environment, the method has the features that: by employing regular ordering of the system connectivity graph, an extremely efficient preconditioner can be derived from the 'range space metric', as opposed to the system coefficient matrix; because of the effectiveness of the preconditioner, preliminary studies indicate that the method can achieve performance rates that depend linearly upon the number of substructures, hence the title 'Order N'; and the method is non-assembling. Furthermore, the approach is promising as a potential parallel processing algorithm in that the method exhibits a fine parallel granularity suitable for a wide collection of combinations of physical system topologies/computer architectures; and the method is easily load balanced among processors, and does not rely upon system topology to induce parallelism.

Kurdila, Andrew J.↗

Structural Integrity of a Wind Tunnel Balance

The National Aeronautics and Space Administration (NASA) Langley Research Center (LaRC) has been designing strain-gage balances for utilization in wind tunnels since its inception. The utilization of balances span over a wide variety of aerodynamic tests. A force balance is an inherently critically stressed component due to the requirements of measurement sensitivity. Research and analyses are done in order to investigate the structural integrity of the balances as well as developing an understanding of their performance in order to enhance their capability. Maximum loading occurs when all 6 components of the loads are applied simultaneously with their maximum value allowed (limit load). This circumstance normally does not occur in the wind tunnel. However, if it occurs, is the balance capable of handling the loads with an acceptable factor of safety? LaRC Balance 1621 was modeled and meshed in PATRAN for analysis in NASTRAN. For a complete analysis, it is necessary to consider all the load cases as well as use dense mesh near all the edges. Because of computer limitations, it is not possible to have one model with the dense mesh near all edges. In the present study, a dense mesh is limited to the surface corners where the cage and axial sections meet. Four different load combinations are used for the current analysis. Linear analysis is performed for each load case. In the case where the stress value is above linear elastic region, it is necessary to perform nonlinear analysis. It is also important to investigate the variables limiting the structural integrity of the balances. In order to investigate the possibility of modifying the existing balances to enhance the structural integrity, some modifications are done on this balance. The structural integrity of the balance after modification is investigated.

Karkehabadi, R.↗

Miniature Remote Deadweight Calibrator

Miniature, computer-controlled deadweight calibrator developed to calibrate remotely force transducer located in cryogenic chamber. Used with microcomputer, calibration system automatically applies deadweight loads to skin-friction balance, records and reduces data, and prints out results. Actuator controller and interlocking-weight system key elements in precise, automatic transducer calibrator. Designed for full-scale load of 1,000 mg. Concept extended to accommodate other full-scale load ranges.

Supplee, Frank H., Jr.↗

Accurate modeling of parallel scientific computations

Scientific codes are usually parallelized by partitioning a grid among processors. To achieve top performance it is necessary to partition the grid so as to balance workload and minimize communication/synchronization costs. This problem is particularly acute when the grid is irregular, changes over the course of the computation, and is not known until load time. Critical mapping and remapping decisions rest on the ability to accurately predict performance, given a description of a grid and its partition. This paper discusses one approach to this problem, and illustrates its use on a one-dimensional fluids code. The models constructed are shown to be accurate, and are used to find optimal remapping schedules.

Nicol, David M.↗

New Approach for the Assessment of Interactions of a Strain-Gage Balance

A new approach was developed to assess interactions of a strain-gage balance. The approach compares interactions observed during a calibration of the balance with interactions observed during the application of check loads. The approach assumes that alignment errors of the calibration loads are negligible. In that case, interactions observed during the calibration can be used as a reference for the assessment of interactions of check loads. The new approach first fits interactions of a single-component load of the calibration data using simple regression models. Then, interactions of the check loads are predicted by using the check load values as inputs for the regression models of the interactions. In the next step, absolute values of differences between predicted and observed interactions of the check loads are computed. Afterwards, the largest differences are compared with empirical thresholds to assess how well the interactions of the check load data agree with the interactions of the calibration data. Recently obtained data sets of a six-component force balance are used to illustrate the application of the new approach. It is also demonstrated how knowledge of balance design characteristics can benefit the assessment of interactions.

wind tunnel test↗

Possibility of saturnian synchrotron radiation

Synchrotron spectra were computed for a hypothetical Saturn radiation belt. A monoenergetic distribution of relativistic electrons was assumed to exist in a toroidal volume in a dipole magnetic field aligned with the rotation axis. When synchrotron emission is added to the thermal component implied by observations at wavelengths shorter than 50 cm, the upturn in the total emission spectrum may be gradual to sharp, depending on the critical wavelength and the number density of radiating electrons. With the constraint of the pressure balance equation, the ranges of magnetic field intensity, and electron energy, and density were tabulated and nomographically illustrated. For several estimates of the magnetic field loading factor, the minimum surface magnetic field intensity and minimum electron number density were computed.

Luthey, J. L.↗

High-Performance Parallel Analysis of Coupled Problems for Aircraft Propulsion

This research program dealt with the application of high-performance computing methods to the numerical simulation of complete jet engines. The program was initiated in January 1993 by applying two-dimensional parallel aeroelastic codes to the interior gas flow problem of a bypass jet engine. The fluid mesh generation, domain decomposition and solution capabilities were successfully tested. Attention was then focused on methodology for the partitioned analysis of the interaction of the gas flow with a flexible structure and with the fluid mesh motion driven by these structural displacements. The latter is treated by a ALE technique that models the fluid mesh motion as that of a fictitious mechanical network laid along the edges of near-field fluid elements. New partitioned analysis procedures to treat this coupled three-component problem were developed during 1994 and 1995. These procedures involved delayed corrections and subcycling, and have been successfully tested on several massively parallel computers, including the iPSC-860, Paragon XP/S and the IBM SP2. For the global steady-state axisymmetric analysis of a complete engine we have decided to use the NASA-sponsored ENG10 program, which uses a regular FV-multiblock-grid discretization in conjunction with circumferential averaging to include effects of blade forces, loss, combustor heat addition, blockage, bleeds and convective mixing. A load-balancing preprocessor tor parallel versions of ENG10 was developed. During 1995 and 1996 we developed the capability tor the first full 3D aeroelastic simulation of a multirow engine stage. This capability was tested on the IBM SP2 parallel supercomputer at NASA Ames. Benchmark results were presented at the 1196 Computational Aeroscience meeting.

Felippa, C. A.↗

The vehicle design evaluation program - A computer-aided design procedure for transport aircraft

The vehicle design evaluation program is described. This program is a computer-aided design procedure that provides a vehicle synthesis capability for vehicle sizing, external load analysis, structural analysis, and cost evaluation. The vehicle sizing subprogram provides geometry, weight, and balance data for aircraft using JP, hydrogen, or methane fuels. The structural synthesis subprogram uses a multistation analysis for aerodynamic surfaces and fuselages to develop theoretical weights and geometric dimensions. The parts definition subprogram uses the geometric data from the structural analysis and develops the predicted fabrication dimensions, parts material raw stock buy requirements, and predicted actual weights. The cost analysis subprogram uses detail part data in conjunction with standard hours, realization factors, labor rates, and material data to develop the manufacturing costs. The program is used to evaluate overall design effects on subsonic commercial type aircraft due to parameter variations.

Oman, B. H.↗

Automatic partitioning of unstructured meshes for the parallel solution of problems in computational mechanics

Most of the recently proposed computational methods for solving partial differential equations on multiprocessor architectures stem from the 'divide and conquer' paradigm and involve some form of domain decomposition. For those methods which also require grids of points or patches of elements, it is often necessary to explicitly partition the underlying mesh, especially when working with local memory parallel processors. In this paper, a family of cost-effective algorithms for the automatic partitioning of arbitrary two- and three-dimensional finite element and finite difference meshes is presented and discussed in view of a domain decomposed solution procedure and parallel processing. The influence of the algorithmic aspects of a solution method (implicit/explicit computations), and the architectural specifics of a multiprocessor (SIMD/MIMD, startup/transmission time), on the design of a mesh partitioning algorithm are discussed. The impact of the partitioning strategy on load balancing, operation count, operator conditioning, rate of convergence and processor mapping is also addressed. Finally, the proposed mesh decomposition algorithms are demonstrated with realistic examples of finite element, finite volume, and finite difference meshes associated with the parallel solution of solid and fluid mechanics problems on the iPSC/2 and iPSC/860 multiprocessors.

Farhat, Charbel↗

Parallel Computing Strategies for Irregular Algorithms

Parallel computing promises several orders of magnitude increase in our ability to solve realistic computationally-intensive problems, but relies on their efficient mapping and execution on large-scale multiprocessor architectures. Unfortunately, many important applications are irregular and dynamic in nature, making their effective parallel implementation a daunting task. Moreover, with the proliferation of parallel architectures and programming paradigms, the typical scientist is faced with a plethora of questions that must be answered in order to obtain an acceptable parallel implementation of the solution algorithm. In this paper, we consider three representative irregular applications: unstructured remeshing, sparse matrix computations, and N-body problems, and parallelize them using various popular programming paradigms on a wide spectrum of computer platforms ranging from state-of-the-art supercomputers to PC clusters. We present the underlying problems, the solution algorithms, and the parallel implementation strategies. Smart load-balancing, partitioning, and ordering techniques are used to enhance parallel performance. Overall results demonstrate the complexity of efficiently parallelizing irregular algorithms.

Biswas, Rupak↗