Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “sparse matrix factorization”

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.

32 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Novel Application of Machine Learning Techniques for Rapid Source Apportionment of Aerosol Mass Spectrometer Datasets

In this work, we apply machine learning approaches sparse multinomial logistic regression to classify aerosol mass spectrometer (AMS) unit mass resolution (UMR) data followed by an ensemble regression technique for source apportionment of organic aerosols (OA). The classifier was trained on 60 well characterized laboratory and positive matrix factorization (PMF) deconvolved reference spectra to identify eight OA types. These include four laboratory-derived secondary organic aerosol (SOA) spectra, which include isoprene photooxidation SOA, isoprene epoxydiols (IEPOX) SOA, a monoterpene SOA type that includes a-pinene and ß-pinene SOA, and aromatic SOA from oxidation of naphthalene and m-xylene precursors, as well as PMF deconvolved spectra for three primary organic aerosol (POA) types, namely, hydrocarbon-like organic aerosol (HOA), biomass burning organic aerosol (BBOA), and cooking OA (COA), and a more oxidized oxygenated OA type (MO-OOA). A 5-fold cross-validation strategy, repeated 10 times, was used to assess the classifier’s performance. The classifier had high classification accuracy for COA, aromatic SOA, and isoprene SOA spectra but incorrectly classified ~9% by number of MO-OOA spectra as BBOA, 12% of BBOA spectra as HOA (and vice versa), and 18% of IEPOX-SOA spectra as aromatic SOA. Next, an ensemble regression model was trained on an artificially generated dataset consisting of mixtures of different OA types to assess its ability to predict fractional mass abundances from classification probabilities of various OA species obtained from the multinomial logistic regression classifier trained on the reference spectra. Ultimately, the proposed approach was applied for source apportionment of aircraft-based AMS measurements of OA UMR spectra during the HI-SCALE field campaign. On two representative days (May 6th and 18th, 2016), the algorithm determined that ~50-60% of OA by mass was MO-OOA, which represented a highly aged organic aerosol mixture from different sources. On both days, BBOA was determined to contribute less than 10% to OA by mass. However, on May 18th, the aromatic SOA fraction was higher compared to that on May 6th. The proposed approach is capable of rapidly analyzing AMS data in real time, making it suitable for applications where rapid source apportionment of AMS OA spectra is desirable.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Novel Solver Algorithms for Nearly Singular Linear Systems Arising in Combustion Modelling

Direct Numerical Simulations of realistic combustion devices are extremely challenging due to the wide separation of scales in the simulation, for example an internal combustion (IC) engine chamber, and the flame thickness of a high-pressure flame. The PeleLMeX solver uses adaptive mesh refinement (AMR) to evolve multi-species reacting flows in the low Mach number limit at the Exascale and relies on an embedded boundary (EB) approach to represent complex geometries. In that framework, the EB geometries often give rise to very small cut-cells along the boundary, which translate into extreme ill-conditioning of the pressure-projection, with eigenvalues that span 15-16 orders of magnitude. In this talk, we focus on the case of a typical IC piston bowl geometry for which we present on a novel approach towards solving these nearly singular linear systems with ILU-based, C-AMG smoothers on massively parallel architectures. In particular, we use scaling and equilibration algorithms to handle the non-normality of the upper triangular factors. This enables us to approximate the highly sequential triangular solve algorithm, embedded in the AMG smoothing-solve phase, with Jacobi iterations. This approximation can be written as a convergent Neumann series whose terms are composed of highly parallel sparse matrix vector multiplications. The result is an algorithm that substantially decreases setup and solve time, compared to state-of-the-art, for these challenging linear systems.

combustion modelling↗
DOE OSTI2024

Batched sparse direct solver design and evaluation in SuperLU_DIST

Over the course of interactions with various application teams, the need for batched sparse linear algebra functions has emerged in order to make more efficient use of the GPUs for many small and sparse linear algebra problems. In this paper, we present our recent work on a batched sparse direct solver for GPUs. The sparse LU factorization is computed by the levels of the elimination tree, leveraging the batched dense operations at each level and a new batched Scatter GPU kernel. The sparse triangular solve is computed by the level sets of the directed acyclic graph (DAG) of the triangular matrix. Batched operations overcome the large overhead associated with launching many small kernels. For medium sized matrix batches with not-so-small bandwidth, using an NVIDIA A100 GPU, our new batched sparse direct solver is orders of magnitude faster than a batched banded solver and uses less than one-tenth of the memory.

Boukaram, Wajih↗
DOE OSTI2024

Accuracy optimized neural networks do not effectively model optic flow tuning in brain area MSTd

Accuracy-optimized convolutional neural networks (CNNs) have emerged as highly effective models at predicting neural responses in brain areas along the primate ventral stream, but it is largely unknown whether they effectively model neurons in the complementary primate dorsal stream. We explored how well CNNs model the optic flow tuning properties of neurons in dorsal area MSTd and we compared our results with the Non-Negative Matrix Factorization (NNMF) model, which successfully models many tuning properties of MSTd neurons. To better understand the role of computational properties in the NNMF model that give rise to optic flow tuning that resembles that of MSTd neurons, we created additional CNN model variants that implement key NNMF constraints – non-negative weights and sparse coding of optic flow. While the CNNs and NNMF models both accurately estimate the observer's self-motion from purely translational or rotational optic flow, NNMF and the CNNs with nonnegative weights yield substantially less accurate estimates than the other CNNs when tested on more complex optic flow that combines observer translation and rotation. Despite its poor accuracy, NNMF gives rise to tuning properties that align more closely with those observed in primate MSTd than any of the accuracy-optimized CNNs. This work offers a step toward a deeper understanding of the computational properties and constraints that describe the optic flow tuning of primate area MSTd.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

A two-level GPU-accelerated incomplete LU preconditioner for general sparse linear systems

This paper presents a parallel preconditioning approach based on incomplete LU (ILU) factorizations in the framework of Domain Decomposition (DD) for general sparse linear systems. We focus on distributed memory parallel architectures, specifically, those that are equipped with graphic processing units (GPUs). In addition to block-Jacobi, we present general purpose two-level ILU Schur complement-based approaches, where different strategies are presented to solve the coarse-level reduced system. These strategies are combined with modified ILU methods in the construction of the coarse-level operator, in order to effectively remove smooth errors by targeting an algebraically smooth vector. We leverage available GPU-based sparse matrix kernels to accelerate the setup and the solve phases of the proposed ILU preconditioner. We evaluate the efficiency of the proposed methods as a smoother for algebraic multigrid (AMG) and as a preconditioner for Krylov subspace methods on challenging anisotropic diffusion problems and a collection of general sparse matrices.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Randomized Algorithms for Symmetric Nonnegative Matrix Factorization

Symmetric Nonnegative Matrix Factorization (SymNMF) is a technique in data analysis and machine learning that approximates a matrix with a product of a nonnegative, low-rank matrix and it transpose. To design faster and more scalable algorithms for SymNMF we develop two randomized algorithms for its computation. The first method uses randomized matrix sketching to compute an initial low-rank approximation to the input matrix and proceeds to uses this as a low-rank input to rapidly compute a SymNMF. The second methods uses randomized leverage score sampling to approximately solve constrained least squares problems. Many successful methods for SymNMF rely on (approximately) solving sequences of constrained least squares problems. Here, we prove theoretically that leverage score sampling can approximately solve constrained least squares problems to e-accuracy. Finally we demonstrate both methods work in practice by applying them to graph clustering tasks on large real world data sets. These experiments show that our methods approximately maintain solution quality and achieve significant speed ups for both large dense and large sparse problems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Learning an Algebriac Multrigrid Interpolation Operator Using a Modified GraphNet Architecture

This work, building on previous efforts, develops a suite of new graph neural network machine learning architectures that generate data-driven prolongators for use in Algebraic Multigrid (AMG). Algebraic Multigrid is a powerful and common technique for solving large, sparse linear systems. Its effectiveness is problem dependent and heavily depends on the choice of the prolongation operator, which interpolates the coarse mesh results onto a finer mesh. Previous work has used recent developments in graph neural networks to learn a prolongation operator from a given coefficient matrix. In this paper, we expand on previous work by exploring architectural enhancements of graph neural networks. A new method for generating a training set is developed which more closely aligns to the test set. Asymptotic error reduction factors are compared on a test suite of 3-dimensional Poisson problems with varying degrees of element stretching. Results show modest improvements in asymptotic error factor over both commonly chosen baselines and learning methods from previous work.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Improved Evaluation of Large Network Matrices for Linear Power Flow Within Optimization Problems: Preprint

This work discusses methods for evaluating the Power Transfer Distribution Factor (PTDF) and Line Outage Distribution Factor (LODF) matrices by employing sparse linear algebra for large-scale computing applications. These matrices are critical in many power systems applications, such as the Unit Commitment Problem (UC), pre- and post-contingency power flow analysis, and transmission expansion. These matrices are typically dense, which means they require a significant amount of time and memory to be computed for large networks. However, by analyzing the structure of the matrices and their computation method, it is possible to use reduced memory methods based on sparse matrix operations. This paper shows that sparse linear algebra algorithms are faster and require less memory and time than traditional dense approaches. Additionally, we explore the effect of matrix sparsification by eliminating trailing digits on power flow calculations.

ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2024

Improved Evaluation of Large Network Matrices for Linear Power Flow Within Optimization Problems

This work presents methods for evaluating the Power Transfer Distribution Factor (PTDF) and Line Outage Distribution Factor (LODF) matrices by employing sparse linear algebra for large-scale computing applications. These matrices play a critical role in many power system applications, such as the Unit Commitment Problem (UC), pre- and post-contingency power flow analysis, and transmission expansion. These matrices are typically dense, which means they require a significant amount of time and memory to be computed for large networks. However, by analyzing the structure of the matrices and their computation method, it is possible to use reduced memory methods based on sparse matrix operations. This paper shows that sparse linear algebra algorithms are faster and require less memory and time than traditional dense approaches. Additionally, we explore the effect of matrix sparsification by eliminating trailing digits on power flow calculations.

large scale↗
DOE OSTI2023

Newly Released Capabilities in the Distributed-Memory SuperLU Sparse Direct Solver

We present the new features available in the recent release of SuperLU_DIST, Version 8.1.1. SuperLU_DIST is a distributed-memory parallel sparse direct solver. The new features include (1) a 3D communication-avoiding algorithm framework that trades off inter-process communication for selective memory duplication, (2) multi-GPU support for both NVIDIA GPUs and AMD GPUs, and (3) mixed-precision routines that perform single-precision LU factorization and double-precision iterative refinement. Apart from the algorithm improvements, we also modernized the software build system to use CMake and Spack package installation tools to simplify the installation procedure. Throughout the article, we describe in detail the pertinent performance-sensitive parameters associated with each new algorithmic feature, show how they are exposed to the users, and give general guidance of how to set these parameters. We illustrate that the solver’s performance both in time and memory can be greatly improved after systematic tuning of the parameters, depending on the input sparse matrix and underlying hardware.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Classical Benchmarks for Variational Quantum Eigensolver Simulations of the Hubbard Model

Simulating the Hubbard model is of great interest to a wide range of applications within condensed matter physics, however its solution on classical computers remains challenging in dimensions larger than one. The relative simplicity of this model, embodied by the sparseness of the Hamiltonian matrix, allows for its efficient implementation on quantum computers, and for its approximate solution using variational algorithms such as the variational quantum eigensolver. While these algorithms have been shown to reproduce the qualitative features of the Hubbard model, their quantitative accuracy in terms of producing true ground state energies and other properties, and the dependence of this accuracy on the system size and interaction strength, the choice of variational ansatz, and the degree of spatial inhomogeneity in the model, remains unknown. Here we present a rigorous classical benchmarking study, demonstrating the potential impact of these factors on the accuracy of the variational solution of the Hubbard model on quantum hardware, for systems with up to 32 qubits. We find that even when using the most accurate wavefunction ansätze for the Hubbard model, the error in its ground state energy and wavefunction plateaus for larger lattices, while stronger electronic correlations magnify this issue. Concurrently, spatially inhomogeneous parameters and the presence of off-site Coulomb interactions only have a small effect on the accuracy of the computed ground state energies. Our study highlights the capabilities and limitations of current approaches for solving the Hubbard model on quantum hardware, and we discuss potential future avenues of research.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

High performance sparse multifrontal solvers on modern GPUs

Here, we have ported the numerical factorization and triangular solve phases of the sparse direct solver STRUMPACK to GPU. STRUMPACK implements sparse LU factorization using the multifrontal algorithm, which performs most of its operations in dense linear algebra operations on so-called frontal matrices of various sizes. Our GPU implementation off-loads these dense linear algebra operations, as well as the sparse scatter–gather operations between frontal matrices. For the larger frontal matrices, our GPU implementation relies on vendor libraries such as cuBLAS and cuSOLVER for NVIDIA GPUs and rocBLAS and rocSOLVER for AMD GPUs. For the smaller frontal matrices we developed custom CUDA and HIP kernels to reduce kernel launch overhead. Overall, high performance is achieved by identifying submatrix factorizations corresponding to sub-trees of the multifrontal assembly tree which fit entirely in GPU memory. The multi-GPU setting uses SLATE (Software for Linear Algebra Targeting Exascale) as a modern GPU-aware replacement for ScaLAPACK. On 4 nodes of SUMMIT the code runs ~10X faster when using all 24 V100 GPUs compared to when it only uses the 168 POWER9 cores. On 8 SUMMIT nodes, using 48 V100 GPUs, the sparse solver reaches over 50TFlop/s. Compared to SuperLU, on a single V100, for a set of 17 matrices our implementation is faster for all but one matrix, and is on average 5X (median 4X) faster

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Bit-GraphBLAS: Bit-Level Optimizations of Matrix-Centric Graph Processing on GPU

In the graph data structure like adjacency matrix, the connectivity of two nodes can be sufficiently represented using only 1 bit, but they are generally treated as 32-bit full-precision in state-of-the-art graph frameworks to adopt common sparse format such as CSR. Meanwhile, bit-level parallelism has recently be explored to have high-performance potential and low storage requirement on GPUs with dense bit-tiles. To fill the gap, our solution is a hierarchical storage format that contains the bit-indexing base and dense bit-tile units. Inherently, the granularity of the bit-tile is an essential factor in achieving both storage compression and GPU parallelism. How to find a sweet spot that trades off between avoiding sparsity and exploiting is comprehensively researched in this work. In the experiment, we evaluate the proposed storage format and algorithms on modern generation GPUs, including Pascal and Volta, to figure out critical software co-designs in conjunction with existing hardware-specific optimization.

Chen, Jou-An↗
DOE OSTI2022

A dictionary learning algorithm for compression and reconstruction of streaming data in preset order

There has been an emerging interest in developing and applying dictionary learning (DL) to process massive datasets in the last decade. Many of these efforts, however, focus on employing DL to compress and extract a set of important features from data, while considering restoring the original data from this set a secondary goal. On the other hand, although several methods are able to process streaming data by updating the dictionary incrementally as new snapshots pass by, most of those algorithms are designed for the setting where the snapshots are randomly drawn from a probability distribution. In this paper, we present a new DL approach to compress and denoise massive dataset in real time, in which the data are streamed through in a preset order (instances are videos and temporal experimental data), so at any time, we can only observe a biased sample set of the whole data. Here, our approach incrementally builds up the dictionary in a relatively simple manner: if the new snapshot is adequately explained by the current dictionary, we perform a sparse coding to find its sparse representation; otherwise, we add the new snapshot to the dictionary, with a Gram-Schmidt process to maintain the orthogonality. To compress and denoise noisy datasets, we apply the denoising to the snapshot directly before sparse coding, which deviates from traditional dictionary learning approach that achieves denoising via sparse coding. Compared to full-batch matrix decomposition methods, where the whole data is kept in memory, and other mini-batch approaches, where unbiased sampling is often assumed, our approach has minimal requirement in data sampling and storage: i) each snapshot is only seen once then discarded, and ii) the snapshots are drawn in a preset order, so can be highly biased. Through experiments on climate simulations and scanning transmission electron microscopy (STEM) data, we demonstrate that the proposed approach performs competitively to those methods in data reconstruction and denoising.

97 MATHEMATICS AND COMPUTING↗