Search NASA⌕ Search

SEARCH · Search NASA

Results for “matrix”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

Exploiting Multiple Levels of Parallelism in Sparse Matrix-Matrix Multiplication

Sparse matrix-matrix multiplication (or SpGEMM) is a key primitive for many high-performance graph algorithms as well as for some linear solvers, such as algebraic multigrid. The scaling of existing parallel implementations of SpGEMM is heavily bound by communication. Even though 3D (or 2.5D) algorithms have been proposed and theoretically analyzed in the flat MPI model on Erdös-Rényi matrices, those algorithms had not been implemented in practice and their complexities had not been analyzed for the general case. In this work, we present the first implementation of the 3D SpGEMM formulation that exploits multiple (intranode and internode) levels of parallelism, achieving significant speedups over the state-of-the-art publicly available codes at all levels of concurrencies. We extensively evaluate our implementation and identify bottlenecks that should be subject to further research.

2.5D algorithms↗
DOE OSTI2016

Matrix product operators, matrix product states, and ab initio density matrix renormalization group algorithms

Current descriptions of the ab initio density matrix renormalization group (DMRG) algorithm use two superficially different languages: an older language of the renormalization group and renormalized operators, and a more recent language of matrix product states and matrix product operators. The same algorithm can appear dramatically different when written in the two different vocabularies. Here, we carefully describe the translation between the two languages in several contexts. First, we describe how to efficiently implement the ab initio DMRG sweep using a matrix product operator based code, and the equivalence to the original renormalized operator implementation. Next we describe how to implement the general matrix product operator/matrix product state algebra within a pure renormalized operator-based DMRG code. Finally, we discuss two improvements of the ab initio DMRG sweep algorithm motivated by matrix product operator language: Hamiltonian compression, and a sum over operators representation that allows for perfect computational parallelism. The connections and correspondences described here serve to link the future developments with the past and are important in the efficient implementation of continuing advances in ab initio DMRG and related algorithms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

A projected preconditioned conjugate gradient algorithm for computing many extreme eigenpairs of a Hermitian matrix

We present an iterative algorithm for computing an invariant subspace associated with the algebraically smallest eigenvalues of a large sparse or structured Hermitian matrix A. We are interested in the case in which the dimension of the invariant subspace is large (e.g., over several hundreds or thousands) even though it may still be small relative to the dimension of A. These problems arise from, for example, density functional theory (DFT) based electronic structure calculations for complex materials. The key feature of our algorithm is that it performs fewer Rayleigh–Ritz calculations compared to existing algorithms such as the locally optimal block preconditioned conjugate gradient or the Davidson algorithm. It is a block algorithm, and hence can take advantage of efficient BLAS3 operations and be implemented with multiple levels of concurrency. We discuss a number of practical issues that must be addressed in order to implement the algorithm efficiently on a high performance computer.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2023

Parameterization of Direct and Doorway Processes in R-Matrix Formalism

R-matrix formalism is extended beyond compound nuclear (CN) resonant reactions to include parameterization of direct as well as doorway processes. Direct processes in the R-matrix exterior are parameterized by a unitary matrix that introduces mixing among wave function coefficients of the incoming and outgoing wave function components at the R-matrix channel surface. Doorway processes are parameterized by separating the Hilbert space of the interior R-matrix region into its doorway and CN subspaces, from which doorway state eigenenergies, reduced width amplitudes, and the strengths of their coupling to CN levels appear as new R-matrix parameters. Parameterization of generalized as well as the conventional Reich–Moore approximation for eliminated capture channels in the presence of direct, doorway, and CN processes is presented along with a complex-valued scattering length with contributions from direct, doorway, and CN capture processes. Derivation of Brune’s alternative R-matrix parameters is extended to include doorway states. This work suggests how R-matrix formalism could be extended further by adopting the concepts from related reaction formalisms.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2016

Ceramic matrix composite article and process of fabricating a ceramic matrix composite article

A ceramic matrix composite article and a process of fabricating a ceramic matrix composite are disclosed. The ceramic matrix composite article includes a matrix distribution pattern formed by a manifold and ceramic matrix composite plies laid up on the matrix distribution pattern, includes the manifold, or a combination thereof. The manifold includes one or more matrix distribution channels operably connected to a delivery interface, the delivery interface configured for providing matrix material to one or more of the ceramic matrix composite plies. The process includes providing the manifold, forming the matrix distribution pattern by transporting the matrix material through the manifold, and contacting the ceramic matrix composite plies with the matrix material.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2018

Ceramic matrix composite article and process of fabricating a ceramic matrix composite article

A ceramic matrix composite article and a process of fabricating a ceramic matrix composite are disclosed. The ceramic matrix composite article includes a matrix distribution pattern formed by a manifold and ceramic matrix composite plies laid up on the matrix distribution pattern, includes the manifold, or a combination thereof. The manifold includes one or more matrix distribution channels operably connected to a delivery interface, the delivery interface configured for providing matrix material to one or more of the ceramic matrix composite plies. The process includes providing the manifold, forming the matrix distribution pattern by transporting the matrix material through the manifold, and contacting the ceramic matrix composite plies with the matrix material.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2021

Scattering matrix pole expansions for complex wave numbers in R -matrix theory

In this followup article to Ducru et al., we establish new results on scattering matrix pole expansions for complex wave numbers in R-matrix theory. In the past, two branches of theoretical formalisms emerged to describe the scattering matrix in nuclear physics: R-matrix theory and pole expansions. The two have been quite isolated from one another. Recently, our study of Brune's alternative parametrization of R-matrix theory has shown the need to extend the scattering matrix (and the underlying R-matrix operators) to complex wave numbers. Two competing ways of doing so have emerged from a historical ambiguity in the definitions of the shift S and penetration P functions: the legacy Lane and Thomas's “force closure” approach versus analytic continuation (which is the standard in mathematical physics). The R-matrix community has not yet come to a consensus as to which to adopt for evaluations in standard nuclear data libraries, such as ENDF. Here, in this article, we argue in favor of analytic continuation of R-matrix operators. We bridge R-matrix theory with the Humblet-Rosenfeld pole expansions, and discover new properties of the Siegert-Humblet radioactive poles and widths, including their invariance properties to changes in channel radii a c . We then show that analytic continuation of R-matrix operators preserves important physical and mathematical properties of the scattering matrix—canceling spurious poles and guaranteeing generalized unitarity—while still being able to close channels below thresholds.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2018

A sparse matrix–vector multiplication based algorithm for accurate density matrix computations on systems of millions of atoms

Here we present an efficient sparse matrix–vector (SpMV) based method to compute the density matrix P from a given Hamiltonian in electronic structure computations. Our method is a hybrid approach based on Chebyshev–Jackson approximation theory and matrix purification methods like the second order spectral projection purification (SP2). Recent methods to compute the density matrix scale as O(N) in the number of floating point operations but are accompanied by large memory and communication overhead, and they are based on iterative use of the sparse matrix–matrix multiplication kernel (SpGEMM), which is known to be computationally irregular. In addition to irregularity in the sparse Hamiltonian H, the nonzero structure of intermediate estimates of P depends on products of H and evolves over the course of computation. On the other hand, an expansion of the density matrix P in terms of Chebyshev polynomials is straightforward and SpMV based; however, the resulting density matrix may not satisfy the required constraints exactly. In this paper, we analyze the strengths and weaknesses of the Chebyshev–Jackson polynomials and the second order spectral projection purification (SP2) method, and propose to combine them so that the accurate density matrix can be computed using the SpMV computational kernel only, and without having to store the density matrix P. Our method accomplishes these objectives by using the Chebyshev polynomial estimate as the initial guess for SP2, which is followed by using sparse matrix–vector multiplications (SpMVs) to replicate the behavior of the SP2 algorithm for purification. We demonstrate the method on a tight-binding model system of an oxide material containing more than 3 million atoms. In addition, we also present the predicted behavior of our method when applied to near-metallic Hamiltonians with a wide energy spectrum.

42 ENGINEERING↗
DOE OSTI2023

Residual stresses in Cu matrix composite surface deposits after laser melt injection

Abstract Tungsten carbide particles reinforced metal matrix composite (MMC) coatings can significantly improve surface wear resistance owing to their increased surface hardness. However, the presence of macro‐ and micro‐residual stresses in MMC coatings can have detrimental effects, such as reducing service life. In this study, neutron diffraction was used to determine the residual stresses in spherical fused tungsten carbide (sFTC) reinforced Cu matrix composite surface deposits after laser melt injection. We also developed a thermo‐mechanical coupled finite element model to predict residual stresses. Our findings reveal that sFTC/Cu composite deposits produced with a preheating temperature of 400°C have low residual stresses, with a maximum tensile residual stress of 98 MPa in the Cu matrix on the top surface. In contrast, the sFTC/bronze (CuAl10Ni5Fe4) composite deposit exhibits very high residual stresses, with a maximum tensile residual stress in the Cu matrix on the top surface reaching 651 MPa. These results provide a better understanding of the magnitudes and distributions of residual stresses in sFTC‐reinforced Cu matrix composite surface deposits manufactured via laser melt injection.

Zhang, Xingxing↗
DOE OSTI2017

Solid matrix transformation and tracer addition using molten ammonium bifluoride salt as a sample preparation method for laser ablation inductively coupled plasma mass spectrometry

Solid sampling and analysis methods, such as laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), are challenged by matrix effects and calibration difficulties. Matrix-matched standards for external calibration are seldom available and it is difficult to distribute spikes evenly into a solid matrix as internal standards. While isotopic ratios of the same element can be measured to high precision, matrix-dependent effects in the sampling and analysis process frustrate accurate quantification and elemental ratio determinations. Here we introduce a potentially general solid matrix transformation approach entailing chemical reactions in molten ammonium bifluoride (ABF) salt that enables the introduction of spikes as tracers or internal standards. Proof of principle experiments show that the decomposition of uranium ore in sealed PFA fluoropolymer vials at 230 °C yields, after cooling, new solids suitable for direct solid sampling by LA. When spikes are included in the molten salt reaction, subsequent LA-ICP-MS sampling at several spots indicate that the spikes are evenly distributed, and that U-235 tracer dramatically improves reproducibility in U-238 analysis. Precisions improved from 17% relative standard deviation for U-238 signals to 0.1% for the ratio of sample U-238 to spiked U-235, a factor of over two orders of magnitude. These results introduce the concept of solid matrix transformation (SMT) using ABF, and provide proof of principle for a new method of incorporating internal standards into a solid for LA-ICP-MS. This new approach, SMT-LA-ICP-MS, provides opportunities to improve calibration and quantification in solids based analysis. Finally, looking forward, tracer addition to transformed solids opens up LA-based methods to analytical methodologies such as standard addition, isotope dilution, preparation of matrix-matched solid standards, external calibration, and monitoring instrument drift against external calibration standard

47 OTHER INSTRUMENTATION↗
DOE OSTI2014

Locality and rapidity of the ultra-large elastic deformation of Nb nanowires in a NiTi phase-transforming matrix

This study investigated the elastic deformation behaviour of Nb nanowires embedded in a NiTi matrix. The Nb nanowires exhibited an ultra-large elastic deformation, which is found to be dictated by the martensitic transformation of the NiTi matrix, thus exhibiting unique characteristics of locality and rapidity. These are in clear contrast to our conventional observation of elastic deformations of crystalline solids, which is a homogeneous lattice distortion with a strain rate controlled by the applied strain. The Nb nanowires are also found to exhibit elastic-plastic deformation accompanying the martensitic transformation of the NiTi matrix in the case when the transformation strain of the matrix over-matches the elastic strain limit of the nanowires, or exhibit only elastic deformation in the case of under-matching. Such insight provides an important opportunity for elastic strain engineering and composite design.

composites↗
DOE OSTI2018

Multithreaded sparse matrix-matrix multiplication for many-core and GPU architectures

Sparse matrix-matrix multiplication is a key kernel that has applications in several domains such as scientific computing and graph analysis. Several algorithms have been studied in the past for this foundational kernel. In this work, we develop parallel algorithms for sparse matrix-matrix multiplication with a focus on performance portability across different high performance computing architectures. The performance of these algorithms depend on the data structures used in them. We compare different types of accumulators in these algorithms and demonstrate the performance difference between these data structures. Furthermore, we develop a meta-algorithm, kkSpGEMM, to choose the right algorithm and data structure based on the characteristics of the problem. We show performance comparisons on three architectures and demonstrate the need for the community to develop two phase sparse matrix-matrix multiplication implementations for efficient reuse of the data structures involved.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Where Lower Calcite Abundance Creates More Alteration: Enhanced Rock Matrix Diffusivity Induced by Preferential Dissolution [Where Lower Calcite Abundance Creates More Alteration: Enhanced Rock Matrix Diffusivity Induced by Preferential Carbonate Dissolution]

Fractured rocks are essential for flow, solute transport and energy production in geosystems. Existing studies on mineral reactions in fractured rocks mostly consider single mineral systems where reactions occur at the fracture wall without changing rock matrix properties. This work presents multicomponent reactive transport numerical experiments in a fractured rock from the Brady’s field, a geothermal reservoir at a depth of 1,396 m in the Hot Springs Mountains, Nevada. Initial porosity, permeability, mineral composition (quartz, clay, and calcite), and fracture geometry are based on microscopy characterization and X-ray tomography. The model was calibrated using a CO 2 -saturated water flooding experiment. Three numerical experiments were carried out with the same initial physical properties however different calcite content. Although total dissolved masses are similar among the three cases, abundant calcite (50% (v/v), calcite50) leads to a localized, thick zone of large porosity increase while low calcite content (10% (v/v), calcite10) creates an extended and narrow zone of small porosity increase resulting in surprisingly larger change in effective transport property. After 300 days of dissolution, effective matrix diffusion coefficients increase by 9.9 and 19.6 times in calcite50 and calcite10, respectively, inducing corresponding 2.1 and 3.2 times rise in the slopes of power law tailing, a measure of transport properties. This counterintuitive results suggest that lower abundance of reactive minerals leads to greater alteration in the fractured media. Detailed analysis show that the effective rates of the fast-dissolving calcite are limited by diffusive transport in the altered matrix and the shape of the altered zone. In contrast, the while effective dissolution of slow-dissolving quartz depends on effective diffusion within the entire rock matrix. Calcite dissolution only occurs at the thin altered–unaltered matrix interface of tens of micrometers thickness occupying less than 1% of the total calcite content. In contrast, all quartz are effectively dissolving. Furthermore, this work highlights the importance of mineralogical complexity in determining mineral dissolution and rock matrix property evolution.

58 GEOSCIENCES↗
DOE OSTI2024

Efficient Mixed-Precision Matrix Factorization of the Inverse Overlap Matrix in Electronic Structure Calculations with AI-Hardware and GPUs

In recent years, a new kind of accelerated hardware has gained popularity in the artificial intelligence (AI) community which enables extremely high-performance tensor contractions in reduced precision for deep neural network calculations. In this article, we exploit Nvidia Tensor cores, a prototypical example of such AI-hardware, to develop a mixed precision approach for computing a dense matrix factorization of the inverse overlap matrix in electronic structure theory, S –1 . This factorization of S –1 , written as ZZT = S –1 , is used to transform the general matrix eigenvalue problem into a standard matrix eigenvalue problem. Here we present a mixed precision iterative refinement algorithm where Z is given recursively using matrix–matrix multiplications and can be computed with high performance on Tensor cores. To understand the performance and accuracy of Tensor cores, comparisons are made to GPU-only implementations in single and double precision. Additionally, we propose a nonparametric stopping criteria which is robust in the face of lower precision floating point operations. The algorithm is particularly useful when we have a good initial guess to Z, for example, from previous time steps in quantum-mechanical molecular dynamics simulations or from a previous iteration in a geometry optimization.

36 MATERIALS SCIENCE↗