Absolute x-ray energy calibration over a wide energy range using a diffraction-based iterative method
Abstract not provided
SEARCH · Engineering Papers
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.
Abstract not provided
Undetected soft errors caused by transient bit flips can lead to silent data corruption (SDC), an undesirable outcome where invalid results pass for valid ones. This has motivated the design of soft error detectors to minimize SDCs. However, the detectors have been studied under different contexts, making comparative evaluation difficult. In this paper, we present the first comprehensive evaluation of four online soft error detection techniques in detecting the adverse impact of soft errors on iterative methods. We observe that, across five iterative methods, the detectors studied achieve high but not perfect detection rates. To understand the potential for improved detection, we evaluate a machine-learning based detector that takes as features that are the runtime features observed by the individual detectors to arrive at their conclusions. Our evaluation demonstrates improved but still far from perfect detection accuracy for the machine learning based detectors. This extensive evaluation demonstrates the need for designing error detectors to handle the evolutionary behavior exhibited by iterative solvers.
Here, we consider hybrid deterministic-stochastic iterative algorithms for the solution of large, sparse linear systems. Starting from a convergent splitting of the coefficient matrix, we analyze various types of Monte Carlo acceleration schemes applied to the original preconditioned Richardson (stationary) iteration. We expect that these methods will have considerable potential for resiliency to faults when implemented on massively parallel machines. We also establish sufficient conditions for the convergence of the hybrid schemes, and we investigate different types of preconditioners including sparse approximate inverses. Numerical experiments on linear systems arising from the discretization of partial differential equations are presented.
The asymptotic iteration method is used to construct the exact energy eigenvalues for a Lorentz vector or a Lorentz scalar, and an equally mixed Lorentz vector and Lorentz scalar Coulombic potentials. Highly accurate and rapidly converging ground-state energies for Lorentz vector Coulomb with a Lorentz vector or a Lorentz scalar linear potential, V(r)=-{lambda}{sub 1}/r+krandV(r)=-{lambda}{sub 1}/randW(r)=kr, respectively, are obtained.
This project supported the Copper Mountain Conference on Multigrid and Iterative Methods, held from 2007 to 2015, at Copper Mountain, Colorado. The subject of the Copper Mountain Conference Series alternated between Multigrid Methods in odd-numbered years and Iterative Methods in even-numbered years. Begun in 1983, the Series represents an important forum for the exchange of ideas in these two closely related fields. This report describes the Copper Mountain Conference on Multigrid and Iterative Methods, 2007-2015. Information on the conference series is available at http://grandmaster.colorado.edu/~copper/.
In this paper, the generalized shift-splitting (GSS) iteration method is used to solve nonsymmetric singular saddle point systems with the positive real (1, 1)-block sub-matrix. By a careful analysis of the spectral properties of the corresponding iteration matrix, we investigate the semi-convergence property of the GSS iteration method, give the sharper bounds of the eigenvalues for the GSS iteration method and propose the inexact GSS iteration method, also. Numerical experiments are given to show the correctness of the theoretical analysis and the feasibility of the GSS preconditioner with appropriate parameters.
Here, the concept of transport mediated through the dynamics of 'jumping' particles is used to develop an iterative method for obtaining steady-state solutions to the nonlocal transport equation in two dimensions (2D). The technique is self-adjoint and capable of correctly treating spatially non-uniform, asymmetric systems. An appropriate reduced version of the iteration method is used to compare with results obtained with a self-adjoint 1D transport matrix approach. The transport 'jump' probability distribution functions (pdfs) are based on Lévy α-stable distributions. The technique can handle the entire Lévy α-parameter range from one (Lorentz distributions) to two (Gaussian distributions). Cases with α = 2 (standard diffusion) are used to establish the validity of the iterative method. The capabilities of the iterative method are demonstrated by presenting examples from systems with various source configurations, boundary shapes, boundary conditions and spatial variations in parameters.
Nuclear criticality calculations with Monte Carlo codes are normally done using a power iteration method to obtain the dominant eigenfunction and eigenvalue. In the last few years it has been shown that the power iteration method can be modified to obtain the first two eigenfunctions. This modified power iteration method directly subtracts out the second eigenfunction and thus only powers out the third and higher eigenfunctions. The result is a convergence rate to the dominant eigenfunction being |k{sub 3}|/k{sub 1} instead of |k{sub 2}|/k{sub 1}. One difficulty is that the second eigenfunction contains particles of both positive and negative weights that must sum somehow to maintain the second eigenfunction. Summing negative and positive weights can be done using point detector mechanics, but this sometimes can be quite slow. We show that an approximate cancellation scheme is sufficient to accelerate the convergence to the dominant eigenfunction. A second difficulty is that for some problems the Monte Carlo implementation of the modified power method has some stability problems. We also show that a simple method deals with this in an effective, but ad hoc manner.
A new method for finding proper factorizations of the hyperpower family of iterative methods for computing generalized inverses is proposed. The method is based on numerical optimization of an appropriate least squares problem. An iterative method in a non-normalized form for finding the generalized outer inverses with prescribed range and null space of any given complex matrix is proposed and considered. Computational complexity and convergence of the method are studied. Several numerical experiments are also presented.
Understanding the impact of soft errors on applications can be expensive. Often, it requires an extensive error injection campaign involving numerous runs of the full application in the presence of errors. In this paper, we present a novel approach to arriving at the ground truth--the true impact of an error on the final output--for iterative methods by observing a small number of iterations to learn deviations between normal and error-impacted execution. We develop a machine learning based predictor for three iterative methods to generate ground-truth results without running them to completion for every error injected. We demonstrate that this approach achieves greater accuracy than alternative prediction strategies, including three existing soft error detection strategies. We demonstrate the effectiveness of the ground truth prediction model in evaluating vulnerability and the effectiveness of soft error detection strategies in the context of iterative methods.
In this paper, we present an effective iterative method for computing symplectic SVD-like decomposition for a 2n-by-m rectangular real matrix A. The main purpose here is a block-power iterative method with the ortho-symplectic SR decomposition in the normalization step. We compute an k-block symplectic SVD-like decomposition, namely A{sub k}=S{sub k}Σ{sub k}V{sub k}{sup T} where S{sub k}∈ℝ{sup 2n×2k} is symplectic and V{sub k}∈ℝ{sup m×2k} is orthogonal. For large matrices, we usually have to compute only part of eigenvalues. The main interest of the proposed method is to efficiently compute the desired number of ordered in magnitude eigenvalues of structured matrices.
Monte Carlo transport codes typically estimate the fundamental mode of a nuclear system via a power iteration method. If the eigenvalues of the system are denoted $k$ 1 , $k$ 2 , $k$ 3 ..." with $k$ 1 > $k$ 2 ≥ $k$ 3 ≥ $k$ 4 ··· > 0, and corresponding eigenfunctions Ψ $m$ , it is well known that the Ψ $m$ mode disappears after $n$ iterations as ($k$ $m$ /$k$ 1 ) $n$ . When the dominance ratio $k$ 2 /$k$ 1 is near unity, the convergence to Ψ 1 can be very slow, even if the convergence to $k$ 1 is fast. The convergence to Ψ 1 sometimes can be dramatically improved by estimating Ψ 1 and Ψ 2 . A Monte Carlo implementation of the modified power iteration method requires the neutrons representing Ψ 2 to have both negative and positive weights because Ψ 2 cannot be of one sign. Thus, Ψ 2 is a sum of negatively and positively weighted neutrons. Instead of summing using point detector methods, a new simple binning procedure accelerates convergence to Ψ 1 .
Currently, the dominating constraint in many high performance computing applications is data capacity and bandwidth, in both internode communications and even moreso in intranode data motion. A new approach to address this limitation is to make use of data compression in the form of a compressed data array. Storing data in a compressed data array and converting to standard IEEE-754 types as needed during a computation can reduce the pressure on bandwidth and storage. However, repeated conversions (lossy compression and decompression) introduce additional approximation errors, which need to be shown to not significantly affect the simulation results. Here, we extend recent work that analyzed the error of a single use of compression and decompression of the ZFP compressed data array representation to the case of time-stepping and iterative schemes, where an advancement operator is repeatedly applied in addition to the conversions. We show that the accumulated error for iterative methods involving fixed-point and time evolving iterations is bounded under standard constraints. An upper bound is established on the number of additional iterations required for the convergence of stationary fixed-point iterations. An additional analysis of traditional forward and backward error of stationary iterative methods using ZFP compressed arrays is also presented. The results of several 1D, 2D, and 3D test problems are provided to demonstrate the correctness of the theoretical bounds.
The paper presents iterative methods for calculating the neutron flux distribution in nonlinear problems of nuclear reactor criticality. Algorithms for solving equations for variations in the neutron flux are considered. Convergence of the iterative processes is studied for two nonlinear problems in which macroscopic interaction cross sections are functionals of the spatial neutron distribution. In the first problem, the neutron flux distribution depends on the water coolant density, and in the second one, it depends on the fuel temperature. Simple relationships connecting the vapor content and the temperature with the neutron flux are used.
Soft errors caused by transient bit flips have the potential to significantly impact an application’s behavior. This has motivated the design of an array of techniques to detect, isolate, and correct soft errors using microarchitectural, architectural, compilation-based, or application-level techniques to minimize their impact on the executing application. The first step towards the design of good error detection/correction techniques involves an understanding of an application’s vulnerability to soft errors. In this paper, we present the first comprehensive characterization of the impact of soft errors on the convergence characteristics of six iterative methods using application-level fault injection. In particular, we consider the use of iterative methods to incrementally solve a linear systems of equations, which constitutes the core kernel in many scientific applications. We analyze the impact of soft errors in terms of the type of error (single- vs multi-bit), the distribution and location of bits affected, the data structure and the statement impacted, and variation with time. In addition to understanding the vulnerability of iterative solvers to soft errors, this characterization can aid the design of fault injection campaigns that ensure systematic coverage.
In this paper, the overlapping local/global (OLG) iteration method based on Monte Carlo/p-CMFD calculations is refined in two aspects. One is the consistent use of estimators to generate homogenized scattering cross sections. Another is that the incident or exiting angular interval is divided into multi-angular bins to modulate albedo boundary conditions for local problems. Numerical tests show that, compared to the one angle bin case in a previous study, the four angle bin case shows significantly improved results. (authors)
Block iterative methods are extremely important as smoothers for multigrid methods, as preconditioners for Krylov methods, and as solvers for diagonally dominant linear systems. Developing robust and efficient smoother algorithms suitable for current and evolving GPU and multicore CPU systems is a significant challenge. We address this issue in the case of constant-coefficient stencils arising in the solution of elliptic partial differential equations on structured 3D uniform and adaptively refined block structured grids. Robust, highly parallel implementations of block Jacobi and chaotic block Gauss-Seidel algorithms with exact inversion of the blocks are developed using different parallelization techniques. Experimental results for NVIDIA Fermi/Kepler GPUs and AMD multicore systems are presented.