SEARCH · Search NASA

Results for “sparse direct method”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3

All sources NASA NTRS DOE OSTI

NASA NTRS2016

The Global Drivers of Photosynthesis and Light Use Efficiency Seasonality: A Granger Frequency Causality Analysis

Photosynthesis and light use efficiency (LUE) are major factors in the evolution of the continental carbon cycle due to their contribution to gross primary production (GPP). However, while the drivers of photosynthesis and LUE on a plant or canopy scale can often be identified, significant uncertainties exist when modeling these on a global scale. This is due to sparse observations in regions such as the tropics and the lack of a direct global observation dataset. Although others have attempted to address this issue using correlations (Beer, 2010) or calculating GPP from vegetation indices (Running, 2004), in this study we take a new approach. We combine the statistical method of Granger frequency causality and partial Granger frequency causality with remote sensing data products (including sun-induced fluorescence used as a proxy for GPP) to determine the main environmental drivers of GPP across the globe.

photosynthesis↗

NASA NTRSDate not supplied

Surface Expression of Bed Topography in Greenland and Antarctica

It has long been recognized that major subglacial topographic features beneath the Greenland and Antarctic ice sheets – whether especially prominent or wide – can generate an observable surface expression. Recent advances in digital elevation models (ArcticDEM, REMA) and bed-to-surface transfer theory now permit widespread observation of this phenomena and a more robust interpretation. Hillshading a digital elevation model from the direction of ice flow permits straightforward detection of major surface features. For Greenland, comparison against the growing catalog of airborne radar-sounding data confirms that the large majority of these features are associated with subglacial topography – typically valleys. These observations suggest a better path toward interpolating subglacial topography between sparse radar observations by developing methods that also require fidelity to observed surface relief.

Joe Macgregor↗

NASA NTRS1987

Automated problem scheduling and reduction of synchronization delay effects

It is anticipated that in order to make effective use of many future high performance architectures, programs will have to exhibit at least a medium grained parallelism. A framework is presented for partitioning very sparse triangular systems of linear equations that is designed to produce favorable preformance results in a wide variety of parallel architectures. Efficient methods for solving these systems are of interest because: (1) they provide a useful model problem for use in exploring heuristics for the aggregation, mapping and scheduling of relatively fine grained computations whose data dependencies are specified by directed acrylic graphs, and (2) because such efficient methods can find direct application in the development of parallel algorithms for scientific computation. Simple expressions are derived that describe how to schedule computational work with varying degrees of granularity. The Encore Multimax was used as a hardware simulator to investigate the performance effects of using the partitioning techniques presented in shared memory architectures with varying relative synchronization costs.

Saltz, Joel H.↗

NASA NTRS1981

An empirical investigation of methods for nonsymmetric linear systems

The present investigation is concerned with a comparison of methods for solving linear algebraic systems which arise from finite difference discretizations of the elliptic convection-diffusion equation in a planar region Omega with Dirichlet boundary conditions. Such linear systems are typically of the form Ax = b where A is an N x N sparse nonsymmetric matrix. In a discussion of discretizations, it is assumed that a regular rectilinear mesh of width h has been imposed on Omega. The discretizations considered include central differences, upstream differences, and modified upstream differences. Six methods for solving Ax = b are considered. Three variants of Gaussian elimination have been chosen as representatives of state-of-the-art software for direct methods under different assumptions about pivoting. Three iterative methods are also included.

Sherman, A. H.↗

NASA NTRS1993

Implementing abstract multigrid or multilevel methods

Multigrid methods can be formulated as an algorithm for an abstract problem that is independent of the partial differential equation, domain, and discretization method. In such an abstract setting, problems not arising from partial differential equations can be treated. A general theory exists for linear problems. The general theory was motivated by a series of abstract solvers (Madpack). The latest version was motivated by the theory. Madpack now allows for a wide variety of iterative and direct solvers, preconditioners, and interpolation and projection schemes, including user callback ones. It allows for sparse, dense, and stencil matrices. Mildly nonlinear problems can be handled. Also, there is a fast, multigrid Poisson solver (two and three dimensions). The type of solvers and design decisions (including language, data structures, external library support, and callbacks) are discussed. Based on the author's experiences with two versions of Madpack, a better approach is proposed. This is based on a mixed language formulation (C and FORTRAN + preprocessor). Reasons for not using FORTRAN, C, or C++ (individually) are given. Implementing the proposed strategy is not difficult.

Douglas, Craig C.↗

NASA NTRS2011

Experiments in Reconstructing Twentieth-Century Sea Levels

One approach to reconstructing historical sea level from the relatively sparse tide-gauge network is to employ Empirical Orthogonal Functions (EOFs) as interpolatory spatial basis functions. The EOFs are determined from independent global data, generally sea-surface heights from either satellite altimetry or a numerical ocean model. The problem is revisited here for sea level since 1900. A new approach to handling the tide-gauge datum problem by direct solution offers possible advantages over the method of integrating sea-level differences, with the potential of eventually adjusting datums into the global terrestrial reference frame. The resulting time series of global mean sea levels appears fairly insensitive to the adopted set of EOFs. In contrast, charts of regional sea level anomalies and trends are very sensitive to the adopted set of EOFs, especially for the sparser network of gauges in the early 20th century. The reconstructions appear especially suspect before 1950 in the tropical Pacific. While this limits some applications of the sea-level reconstructions, the sensitivity does appear adequately captured by formal uncertainties. All our solutions show regional trends over the past five decades to be fairly uniform throughout the global ocean, in contrast to trends observed over the shorter altimeter era. Consistent with several previous estimates, the global sea-level rise since 1900 is 1.70 +/- 0.26 mm/yr. The global trend since 1995 exceeds 3 mm/yr which is consistent with altimeter measurements, but this large trend was possibly also reached between 1935 and 1950.

Ray, Richard D.↗

NASA NTRS1994

Performance issues for iterative solvers in device simulation

Due to memory limitations, iterative methods have become the method of choice for large scale semiconductor device simulation. However, it is well known that these methods still suffer from reliability problems. The linear systems which appear in numerical simulation of semiconductor devices are notoriously ill-conditioned. In order to produce robust algorithms for practical problems, careful attention must be given to many implementation issues. This paper concentrates on strategies for developing robust preconditioners. In addition, effective data structures and convergence check issues are also discussed. These algorithms are compared with a standard direct sparse matrix solver on a variety of problems.

Fan, Qing↗

NASA NTRS1991

Lanczos eigensolution method for high-performance computers

The theory, computational analysis, and applications are presented of a Lanczos algorithm on high performance computers. The computationally intensive steps of the algorithm are identified as: the matrix factorization, the forward/backward equation solution, and the matrix vector multiples. These computational steps are optimized to exploit the vector and parallel capabilities of high performance computers. The savings in computational time from applying optimization techniques such as: variable band and sparse data storage and access, loop unrolling, use of local memory, and compiler directives are presented. Two large scale structural analysis applications are described: the buckling of a composite blade stiffened panel with a cutout, and the vibration analysis of a high speed civil transport. The sequential computational time for the panel problem executed on a CONVEX computer of 181.6 seconds was decreased to 14.1 seconds with the optimized vector algorithm. The best computational time of 23 seconds for the transport problem with 17,000 degs of freedom was on the the Cray-YMP using an average of 3.63 processors.

Bostic, Susan W.↗

NASA NTRS1988

Newton solution of inviscid and viscous problems

The application of Newton iteration to inviscid and viscous airfoil calculations is examined. Spatial discretization is performed using upwind differences with split fluxes. The system of linear equations which arises as a result of linearization in time is solved directly using either a banded matrix solver or a sparse matrix solver. In the latter case, the solver is used in conjunction with the nested dissection strategy, whose implementation for airfoil calculations is discussed. The boundary conditions are also implemented in a fully implicit manner, thus yielding quadratic convergence. Complexities such as the ordering of cell nodes and the use of a far field vortex to correct freestream for a lifting airfoil are addressed. Various methods to accelerate convergence and improve computational efficiency while using Newton iteration are discussed. Results are presented for inviscid, transonic nonlifting and lifting airfoils and also for laminar viscous cases.

Venkatakrishnan, V.↗

NASA NTRSDate not supplied

Performance Optimization Methods for a Memory-Bound, Unstructured-Grid CFD Application on Massively Parallel GPU Platforms

Computational performance of the FUN3D unstructured-grid computational fluid dynamics (CFD) application on massively parallel GPU environments is memory-bound and highly dependent upon efficient reads from and atomic updates to the irregular cell-, edge-, and node-based data structures. In this talk, we present recent efforts into optimizing select performance-critical kernels on NVIDIA Tesla V100 and A100 GPUs and AMD CDNA MI100 GPUs. A novel use of L2 cache residency controls and asynchronous loads into on-chip shared memory are explored on the A100 GPU for the sparse iterative solver, which is dominated by mixed-precision, sparse matrix vector multiplication. Demonstrations show that these methods improve global memory bandwidth utilization by 13.5% on the A100 GPU. Several techniques are also presented that use registers and/or shared memory to facilitate array transposition and aggregation which combine to reduce the frequency and increase the cache efficiency of floating-point atomic updates to the irregular data structures. These methods are demonstrated to improve the kernel throughput by nearly 500% on select kernels on the AMD MI100 over atomic updates directly to global memory. Overall, both V100 and A100 GPUs outperformed the MI100 GPU on kernels dominated by double-precision atomic updates; however, the techniques demonstrated here reduced the performance gap and improved the MI100 performance.

GPU CPU unstructured CFD memory↗

NASA NTRS1999

Utilizing Multiple Datasets for Snow Cover Mapping

Snow-cover maps generated from surface data are based on direct measurements, however they are prone to interpolation errors where climate stations are sparsely distributed. Snow cover is clearly discernable using satellite-attained optical data because of the high albedo of snow, yet the surface is often obscured by cloud cover. Passive microwave (PM) data is unaffected by clouds, however, the snow-cover signature is significantly affected by melting snow and the microwaves may be transparent to thin snow (less than 3cm). Both optical and microwave sensors have problems discerning snow beneath forest canopies. This paper describes a method that combines ground and satellite data to produce a Multiple-Dataset Snow-Cover Product (MDSCP). Comparisons with current snow-cover products show that the MDSCP draws together the advantages of each of its component products while minimizing their potential errors. Improved estimates of the snow-covered area are derived through the addition of two snow-cover classes ("thin or patchy" and "high elevation" snow cover) and from the analysis of the climate station data within each class. The compatibility of this method for use with Moderate Resolution Imaging Spectroradiometer (MODIS) data, which will be available in 2000, is also discussed. With the assimilation of these data, the resolution of the MDSCP would be improved both spatially and temporally and the analysis would become completely automated.

Tait, Andrew B.↗

NASA NTRS2007

Using MODIS and GLAS Data to Develop Timber Volume Estimates in Central Siberia

The boreal forest is the Earth's largest terrestrial biome, covering some 12 million km2 and accounting for about one third of this planet's total forest area. Mapping of boreal forest's type, structure parameters and biomass are critical for understanding the boreal forest's significance in the carbon cycle, its response to and impact on global climate change. Ground based forest inventories, have much uncertainty in the inventory data, particularly in remote areas of Siberia where sampling is sparse and/or lacking. In addition, many of the forest inventories that do exist for Siberia are now a decade or more old. Thus, available forest inventories fail to capture the current conditions. Changes in forest structure in a particular forest-type and region can change significantly due to changing environment conditions, and natural and anthropogenic disturbance. Remote sensing methods can potentially overcome these problems. Multispectral sensors can be used to provide vegetation cover maps that show a timely and accurate geographic distribution of vegetation types rather than decade old ground based maps. Lidar sensors can be used to directly obtain measurements that can be used to derive critical forest structure information (e.g., height, density, and volume). These in turn can used to estimate biomass components using allometric equations without having to use out dated forest inventory. Finally, remote sensing data is ideally suited to provide a sampling basis for a rigorous statistical estimate of the variance and error bound on forest structure measures. In this study, new remote sensing methods were applied to develop estimates timber volume using NASA's MODerate resolution Imaging Spectroradiometer (MODIS) and unique waveform data of the geoscience laser altimeter system (GLAS) for a 10 deg x 10 deg area in central Siberia. Using MODIS and GLAS data, maps were produced for cover type and timber volume for 2003, and a realistic variance (error bound) for timber volume was calculated for the study area. In this 'study we used only GLAS footprints that had a slope value of less than 10 deg. This was done to avoid large errors due to the effect of slope on the GLAS models. The method requires the integration of new remote sensing methods with available ground studies of forest timber volume conducted in Russian forests. The results were compared to traditional ground forest inventory methods reported in the literature and to ground truth collected in the study area.

Ranson, K. Jon↗

NASA NTRS1998

Parallel Implicit Algorithms for CFD

The main goal of this project was efficient distributed parallel and workstation cluster implementations of Newton-Krylov-Schwarz (NKS) solvers for implicit Computational Fluid Dynamics (CFD.) "Newton" refers to a quadratically convergent nonlinear iteration using gradient information based on the true residual, "Krylov" to an inner linear iteration that accesses the Jacobian matrix only through highly parallelizable sparse matrix-vector products, and "Schwarz" to a domain decomposition form of preconditioning the inner Krylov iterations with primarily neighbor-only exchange of data between the processors. Prior experience has established that Newton-Krylov methods are competitive solvers in the CFD context and that Krylov-Schwarz methods port well to distributed memory computers. The combination of the techniques into Newton-Krylov-Schwarz was implemented on 2D and 3D unstructured Euler codes on the parallel testbeds that used to be at LaRC and on several other parallel computers operated by other agencies or made available by the vendors. Early implementations were made directly in Massively Parallel Integration (MPI) with parallel solvers we adapted from legacy NASA codes and enhanced for full NKS functionality. Later implementations were made in the framework of the PETSC library from Argonne National Laboratory, which now includes pseudo-transient continuation Newton-Krylov-Schwarz solver capability (as a result of demands we made upon PETSC during our early porting experiences). A secondary project pursued with funding from this contract was parallel implicit solvers in acoustics, specifically in the Helmholtz formulation. A 2D acoustic inverse problem has been solved in parallel within the PETSC framework.

Keyes, David E.↗

NASA NTRS1986

Investigation of advanced fault insertion and simulator methods

The cooperative agreement partly supported research leading to the open-literature publication cited. Additional efforts under the agreement included research into fault modeling of semiconductor devices. Results of this research are presented in this report which is summarized in the following paragraphs. As a result of the cited research, it appears that semiconductor failure mechanism data is abundant but of little use in developing pin-level device models. Failure mode data on the other hand does exist but is too sparse to be of any statistical use in developing fault models. What is significant in the failure mode data is that, unlike classical logic, MSI and LSI devices do exhibit more than 'stuck-at' and open/short failure modes. Specifically they are dominated by parametric failures and functional anomalies that can include intermittent faults and multiple-pin failures. The report discusses methods of developing composite pin-level models based on extrapolation of semiconductor device failure mechanisms, failure modes, results of temperature stress testing and functional modeling. Limitations of this model particularly with regard to determination of fault detection coverage and latency time measurement are discussed. Indicated research directions are presented.

Dunn, W. R.↗

NASA NTRS1999

Middle Atmosphere Transport Properties of Assimilated Datasets

One of the most compelling reasons for performing data assimilation in the middle atmosphere is to obtain global, balanced datasets for studies of trace gas transport and chemistry. This is a major motivation behind the Goddard Earth observation System-Data Assimilation System (GEOS-DAS). Previous studies have shown that while this and other data assimilation systems can generally obtain good estimates of the extratropical rotational velocity field, the divergent part of the dynamical field is deficient; this impacts the "residual circulation" and leads to spurious trace gas transport on seasonal and interannual timescales. These problems are impacted by the quality and the method of use of the observational data and by deficiencies in the atmospheric general circulation model. Whichever the cause at any place and time, the "solution" is to introduce non-physical forcing terms into the system (the so-called incremental analysis updates); these can directly (thermal) or indirectly (mechanical) affect the residual circulation. This paper will illustrate how the divergent circulation is affected by deficiencies in both observations and models. Theoretical considerations will be illustrated with examples from the GEOS-DAS and from simplified numerical experiments. These are designed to isolate known problems, such as the inability of models to sustain a quasi-biennial oscillation and sparse observational constraints on tropical dynamics, or radiative inconsistencies in the presence of volcanic aerosols.

Pawson, Steven↗

NASA NTRS1999

Middle Atmospheric Transport Properties of Assimilated Datasets

Pawson, Steven↗

NASA NTRS2017

The Effectiveness of Shrouding on Reducing Meshed Spur Gear Power Loss - Test Results

Gearbox efficiency is reduced at high rotational speeds due to windage drag and viscous effects on rotating, meshed gear components. A goal of NASA aeronautics rotorcraft research is aimed at propulsion technologies that improve efficiency while minimizing vehicle weight. Specifically, reducing power losses to rotorcraft gearboxes would allow gains in areas such as vehicle payload, range, mission type, and fuel consumption. To that end, a gear windage rig has been commissioned at NASA Glenn Research Center to measure windage drag on gears and to test methodologies to mitigate windage power losses. One method used in rotorcraft gearbox design attempts to reduce gear windage power loss by utilizing close clearance walls to enclose the gears in both the axial and radial directions. The close clearance shrouds result in reduced drag on the gear teeth, and reduced power loss. For meshed spur gears, the shrouding takes the form of metal side plates and circumferential metal sectors. Variably positioned axial and radial shrouds are incorporated in the NASA rig to study the effect of shroud clearance on gearbox power loss. A number of researchers have given experimental and analytical results for single spur gears, with and without shrouding. Shrouded meshed spur gear test results are sparse in the literature. Windage tests were run at NASA Glenn using meshed spur gears at four shroud configurations: unshrouded, shrouded (max. axial, max radial), and two intermediate shrouding conditions. Results are compared to available meshed spur gear power loss data analyses as well as single spur gear data/analyses. Recommendations are made for future work.

power loss↗

NASA NTRS2017

The Effectiveness of Shrouding on Reducing Meshed Spur Gear Power Loss Test Results

Gearbox efficiency is reduced at high rotational speeds due to windage drag and viscous effects on rotating, meshed gear components. A goal of NASA aeronautics rotorcraft research is aimed at propulsion technologies that improve efficiency while minimizing vehicle weight. Specifically, reducing power losses to rotorcraft gearboxes would allow gains in areas such as vehicle payload, range, mission type, and fuel consumption. To that end, a gear windage rig has been commissioned at NASA Glenn Research Center to measure windage drag on gears and to test methodologies to mitigate windage power losses. One method used in rotorcraft gearbox design attempts to reduce gear windage power loss by utilizing close clearance walls to enclose the gears in both the axial and radial directions. The close clearance shrouds result in reduced drag on the gear teeth and reduced power loss. For meshed spur gears, the shrouding takes the form of metal side plates and circumferential metal sectors. Variably positioned axial and radial shrouds are incorporated in the NASA rig to study the effect of shroud clearance on gearbox power loss. A number of researchers have given experimental and analytical results for single spur gears, with and without shrouding. Shrouded meshed spur gear test results are sparse in the literature. Windage tests were run at NASA Glenn using meshed spur gears at four shroud configurations: unshrouded, shrouded (max. axial, max. radial), and two intermediate shrouding conditions. Results are compared to available meshed spur gear power loss data analyses as well as single spur gear data analyses.

windage↗