Search NASA⌕ Search

SEARCH · Search NASA

Results for “dense LU 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.

33 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Optimizing Communication in 2D Grid-Based MPI Applications at Exascale

The new reality of exascale computing faces many challenges in achieving optimal performance on large numbers of nodes. A key challenge is the efficient utilization of the message-passing interface (MPI), a critical component for process communication. This paper explores communication optimization strategies to harness the GPU-accelerated architectures of these supercomputers. We focus on MPI applications where processors form a two-dimensional process grid, a common arrangement in applications involving dense matrix operations. This configuration offers a unique opportunity to implement innovative strategies to improve performance and maintain effective load distribution. We study two applications— Dist-FW (Apsp:all-pair-shortest-path) and HPL-MxP (LU factorization with Mixed precision)—on two accelerated systems: Summit (IBM Power and NVIDIA V100) and Frontier (AMD EPYC and MI250X). These supercomputers are operated by the Oak Ridge Leadership Computing Facility (OLCF) and are currently ranked #1 and #5 on the Top500 list. We show how to scale up both applications to exascale levels and tackle the MPI challenges related to implementation, synchronization, and performance. We also compare the performance of several communication strategies at an unprecedented scale. Accurately predicting application performance becomes crucial for cost reduction as the computational scale grows. To address this, we suggest a hyperbolic model as a better alternative to the traditional one-sided asymptotic model for predicting future application performance at such large scales.

Lu, Hao↗
DOE OSTI2019

Distributed-memory lattice H-matrix factorization

We parallelize the LU factorization of a hierarchical low-rank matrix (H-matrix) on a distributed-memory computer. This is much more difficult than the H-matrix-vector multiplication due to the dataflow of the factorization, and it is much harder than the parallelization of a dense matrix factorization due to the irregular hierarchical block structure of the matrix. Block low-rank (BLR) format gets rid of the hierarchy and simplifies the parallelization, often increasing concurrency. However, this comes at a price of losing the near-linear complexity of the H-matrix factorization. In this work, we propose to factorize the matrix using a “lattice H-matrix” format that generalizes the BLR format by storing each of the blocks (both diagonals and off-diagonals) in the H-matrix format. These blocks stored in the H-matrix format are referred to as lattices. Thus, this lattice format aims to combine the parallel scalability of BLR factorization with the near-linear complexity of H-matrix factorization. We first compare factorization performances using the H-matrix, BLR, and lattice H-matrix formats under various conditions on a shared-memory computer. Our performance results show that the lattice format has storage and computational complexities similar to those of the H-matrix format, and hence a much lower cost of factorization than BLR. In conclusion, we then compare the BLR and lattice H-matrix factorization on distributed-memory computers. Our performance results demonstrate that compared with BLR, the lattice format with the lower cost of factorization may lead to faster factorization on the distributed-memory computer.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

A graphics processing unit accelerated sparse direct solver and preconditioner with block low rank compression

We present the GPU implementation efforts and challenges of the sparse solver package STRUMPACK. The code is made publicly available on github with a permissive BSD license. STRUMPACK implements an approximate multifrontal solver, a sparse LU factorization which makes use of compression methods to accelerate time to solution and reduce memory usage. Multiple compression schemes based on rank-structured and hierarchical matrix approximations are supported, including hierarchically semi-separable, hierarchically off-diagonal butterfly, and block low rank. Here, in this paper, we present the GPU implementation of the block low rank (BLR) compression method within a multifrontal solver. Our GPU implementation relies on highly optimized vendor libraries such as cuBLAS and cuSOLVER for NVIDIA GPUs, rocBLAS and rocSOLVER for AMD GPUs and the Intel oneAPI Math Kernel Library (oneMKL) for Intel GPUs. Additionally, we rely on external open source libraries such as SLATE (Software for Linear Algebra Targeting Exascale), MAGMA (Matrix Algebra on GPU and Multi-core Architectures), and KBLAS (KAUST BLAS). SLATE is used as a GPU-capable ScaLAPACK replacement. From MAGMA we use variable sized batched dense linear algebra operations such as GEMM, TRSM and LU with partial pivoting. KBLAS provides efficient (batched) low rank matrix compression for NVIDIA GPUs using an adaptive randomized sampling scheme. The resulting sparse solver and preconditioner runs on NVIDIA, AMD and Intel GPUs. Interfaces are available from PETSc, Trilinos and MFEM, or the solver can be used directly in user code. We report results for a range of benchmark applications, using the Perlmutter system from NERSC, Frontier from ORNL, and Aurora from ALCF. For a high frequency wave equation on a regular mesh, using 32 Perlmutter compute nodes, the factorization phase of the exact GPU solver is about 6.5× faster compared to the CPU-only solver. The BLR-enabled GPU solver is about 13.8× faster than the CPU exact solver. For a collection of SuiteSparse matrices, the STRUMPACK exact factorization on a single GPU is on average 1.9× faster than NVIDIA’s cuDSS solver.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

A fast, memory efficient and robust sparse preconditioner based on a multifrontal approach with applications to finite‐element matrices

Summary In this article, we introduce a fast, memory efficient and robust sparse preconditioner that is based on a direct factorization scheme for sparse matrices arising from the finite‐element discretization of elliptic partial differential equations. We use a fast (but approximate) multifrontal approach as a preconditioner and use an iterative scheme to achieve a desired accuracy. This approach combines the advantages of direct and iterative schemes to arrive at a fast, robust, and accurate preconditioner. We will show that this approach is faster (∼2×) and more memory efficient (∼2–3×) than a conventional direct multifrontal approach. Furthermore, we will demonstrate that this preconditioner is both faster and more effective than other preconditioners such as the incomplete LU preconditioner. Specific speedups depend on the matrix size and improve as the size of the matrix increases. The preconditioner can be applied to both structured and unstructured meshes in a similar manner. We build on our previous work and utilize the fact that dense frontal and update matrices, in the multifrontal algorithm, can be represented as hierarchically off‐diagonal low‐rank matrices. Using this idea, we replace all large dense matrix operations in the multifrontal elimination process with O ( N ) hierarchically off‐diagonal low‐rank operations to arrive at a faster and more memory efficient factorization scheme. We then use this direct factorization method at low accuracies as a preconditioner and apply it to various real‐life engineering test cases. Copyright © 2016 John Wiley & Sons, Ltd.

Aminfar, AmirHossein↗
DOE OSTI2018

The Sliding Wear and Friction Behavior of M50-Graphene Self-Lubricating Composites Prepared by Laser Additive Manufacturing at Elevated Temperature

M50 steel is widely applied to manufacture aircraft bearings where service lives are mainly determined by the friction and wear behaviors. The main purpose of this study is to investigate the tribological behaviors and wear mechanisms of M50-1.5 wt.% graphene composites (MGC) prepared by laser additive manufacturing (LAM) (MGC-LAM) sliding against Si{sub 3}N{sub 4} ball from 25 to 550 °C at 18 N–0.2 m/s. XRD, EPMA, FESEM, and EDS mapping were conducted to understand the major mechanisms leading to the improvement in the sliding behavior of MGC-LAM. The results indicated that MGC-LAM showed the excellent friction and wear performance at 25-550 °C for the lower friction coefficient of 0.16-0.52 and less wear rate of 6.1-9.5 × 10{sup −7} mm{sup 3} N{sup −1} m{sup −1}. Especially at 350 °C, MGC-LAM obtained the best tribological performance (0.16, 6.1 × 10{sup −7}mm{sup 3} N{sup −1} m{sup −1}). It was attributed to the dense coral-like microstructure, as well as the formed surface lubricating structure which is composed of the upper uniform lubricating film with massive graphene and the underneath compacted layer.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2021

A Bi-Layer Dense/Porous Solid Electrolyte Interphase for Enhanced Lithium-Metal Stability

Due to its high theoretical capacity and low electrochemical potential, lithium metal is a highly investigated anode for next-generation high-energy batteries. However, the unstable chemical and topographical heterogeneous surface of lithium gives rise to safety and efficiency concerns that prevent it from being utilized in practical applications. Exposure to electrolyte leads to non-uniform, mixed organic-inorganic solid-electrolyte interphase (SEI) formation, starting a cascading to non-uniform flux distribution, consumption of active materials, and dendrite growth. The SEI is the key feature that will lead to harnessing the capabilities possible with lithium metal. In this work, the formation of a closed-host bi-layer solid electrolyte interphase (SEI) improves the stability of lithium anode. This is successfully realized by forming an interconnected porous LiF-rich artificial SEI in contact with Li metal, and a dense, stable in-situ formed upper layer SEI. The porous layer increases the number of Li/LiF interfaces, which reduces local volume fluctuations and improves Li+ diffusion along these interfaces. Additionally, the tortuous porous structure guides uniform Li+ flux distribution and mechanically suppresses dendrite propagation. The dense upper layer of the SEI accomplishes a closed-host design through reaction with LiNO3 additive and prevents continuous consumption of active materials seen without the additive included in the electrolyte. The duality of a dense top layer with porous bottom layer led to extended cycle life and improved rate performance, evidenced with symmetric cell testing, as well as full cell testing paired with sulfur and LiFePO4 (LFP) cathodes. This work is a good example of a rational design of the SEI, based on comprehensive consideration of various critical factors to improve Li-metal anode stability, and highlights a new pathway to improve cycling and rate performances of Li metal batteries.

25 ENERGY STORAGE↗
DOE OSTI2016

Superconductivity in dense carbon-based materials

Guided by a simple strategy in searching of new superconducting materials we predict that high temperature superconductivity can be realized in classes of high-density materials having strong sp 3 chemical bonding and high lattice symmetry. Here, we examine in detail sodalite carbon frameworks doped with simple metals such as Li, Na, and Al. Though such materials share some common features with doped diamond, their doping level is not limited and the density of states at the Fermi level in them can be as high as that in the renowned MgB 2 . Altogether, with other factors, this boosts the superconducting temperature (T c ) in the materials investigated to higher levels compared to doped diamond. For example, the superconducting T c of sodalite-like NaC 6 is predicted to be above 100 K. This phase and a series of other sodalite-based superconductors are predicted to be metastable phases but are dynamically stable. In owing to the rigid carbon framework of these and related dense carbon-materials, these doped sodalite-based structures could be recoverable as potentially useful superconductors.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2017

Efficient Scalable Parallel Higher Order Direct MoM-SIE Method With Hierarchically Semiseparable Structures for 3-D Scattering

A novel fast scalable parallel algorithm is proposed for the solution of large 3-D scattering problems based on: 1) the double (geometrical and current-approximation) higher order (DHO) method of moments (MoM) in the surface integral equation (SIE) formulation and 2) a direct solver for dense linear systems utilizing hierarchically semiseparable (HSS) structures. Namely, an HSS matrix representation is used for compression, factorization, and solution of the system matrix. In addition, a rank-revealing QR decomposition for memory compression is used, with a stopping criterion in terms of the relative rank tolerance value. One method for geometrical preprocessing of the scatterers based on the cobblestone distance sorting technique is employed in order to enhance the HSS algorithm accuracy and parallelization. Numerical examples show how the accuracy of the DHO HSS-MoM-SIE method is easily controllable by using the relative tolerance for the matrix compression. Furthermore, the examples demonstrate low memory consumption, as well as much faster simulation time, when compared to the direct LU decomposition. The method enables dramatically faster monostatic scattering computations than iterative solvers and reduced number of unknowns when compared to low-order discretizations. Finally, great scalability of the algorithm is demonstrated on more than one thousand processes.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2014

Properties of the remnant clockwise disk of young stars in the galactic center

We present new kinematic measurements and modeling of a sample of 116 young stars in the central parsec of the Galaxy in order to investigate the properties of the young stellar disk. The measurements were derived from a combination of speckle and laser guide star adaptive optics imaging and integral field spectroscopy from the Keck telescopes. Compared to earlier disk studies, the most important kinematic measurement improvement is in the precision of the accelerations in the plane of the sky, which have a factor of six smaller uncertainties (σ ∼ 10 μas yr{sup –2}). We have also added the first radial velocity measurements for eight young stars, increasing the sample at the largest radii (6''-12'') by 25%. We derive the ensemble properties of the observed stars using Monte Carlo simulations of mock data. There is one highly significant kinematic feature (∼20σ), corresponding to the well-known clockwise disk, and no significant feature is detected at the location of the previously claimed counterclockwise disk. The true disk fraction is estimated to be ∼20%, a factor of ∼2.5 lower than previous claims, suggesting that we may be observing the remnant of what used to be a more densely populated stellar disk. The similarity in the kinematic properties of the B stars and the O/WR stars suggests a common star formation event. The intrinsic eccentricity distribution of the disk stars is unimodal, with an average value of (e) = 0.27 ± 0.07, which we show can be achieved through dynamical relaxation in an initially circular disk with a moderately top-heavy mass function.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2017

ALMA Maps of Dust and Warm Dense Gas Emission in the Starburst Galaxy IC 5179

We present our high-resolution (0.″15 × 0.″13, ∼34 pc) observations of the CO (6−5) line emission, which probes the warm and dense molecular gas, and the 434 μ m dust continuum emission in the nuclear region of the starburst galaxy IC 5179, conducted with the Atacama Large Millimeter Array (ALMA). The CO (6−5) emission is spatially distributed in filamentary structures with many dense cores and shows a velocity field that is characteristic of a circumnuclear rotating gas disk, with 90% of the rotation speed arising within a radius of ≲150 pc. At the scale of our spatial resolution, the CO (6−5) and dust emission peaks do not always coincide, with their surface brightness ratio varying by a factor of ∼10. This result suggests that their excitation mechanisms are likely different, as further evidenced by the southwest to northeast spatial gradient of both CO-to-dust continuum ratio and Pa- α equivalent width. Within the nuclear region (radius ∼ 300 pc) and with a resolution of ∼34 pc, the CO line flux (dust flux density) detected in our ALMA observations is 180 ± 18 Jy km s{sup −1} (71 ± 7 mJy), which accounts for 22% (2.4%) of the total value measured by Herschel .

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

Sequential surface synthesis of dispersed sub-nanometer iridium on titanium nitride for acidic water oxidation

Maximizing iridium utilization while maintaining high oxygen evolution reaction (OER) performance remains a persistent challenge in acidic water electrolysis. Immobilizing Ir on conductive, acid-stable supports is promising, yet simultaneously achieving sub-nanometer size, high area coverage, and strong electronic coupling is difficult. Here, we report a sequential surface synthesis on titanium nitride (TiN) that yields uniformly distributed sub-nanometer Ir arrays (∼0.7 nm). Our method uses ethylenediaminetetraacetic acid (EDTA) as a temporal scaffold: it chemisorbs to TiN to install dense chelating sites, captures Ir 3+ ions, and confines Ir cluster growth. A subsequent thermal treatment at 500 °C in a reducing atmosphere removes the ligand shell, while preserving ultrasmall particle size and establishing direct Ir–TiN electronic coupling. The optimized catalyst exhibits mixed Ir 0 /Ir x+ coordination with low charge-transfer resistance (R ct = 19.2 Ω), delivering a mass activity of 342 A g Ir −1 at 1.54 V in acidic electrolyte. In situ X-ray absorption spectroscopy reveals irreversible surface oxidation as the primary stability-limiting factor. This stepwise strategy provides a general framework for supported catalysts that maximize precious metal utilization via sub-nanometer dispersion.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Evoking ordered vacancies in metallic nanostructures toward a vacated Barlow packing for high-performance hydrogen evolution

Metallic nanostructures are commonly densely packed into a few packing variants with slightly different atomic packing factors. The structural aspects and physicochemical properties related with the vacancies in such nanostructures are rarely explored because of lack of an effective way to control the introduction of vacancy sites. Highly voided metallic nanostructures with ordered vacancies are however energetically high lying and very difficult to synthesize. Here, we report a chemical method for synthesis of hierarchical Rh nanostructures (Rh NSs) composed of ultrathin nanosheets, composed of hexagonal close-packed structure embedded with nanodomains that adopt a vacated Barlow packing with ordered vacancies. The obtained Rh NSs exhibit remarkably enhanced electrocatalytic activity and stability toward the hydrogen evolution reaction (HER) in alkaline media. Theoretical calculations reveal that the exceptional electrocatalytic performance of Rh NSs originates from their unique vacancy structures, which facilitate the adsorption and dissociation of H 2 O in the HER.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Spatial Simulation and Characterization of Three-Dimensional Fractures in Gejiu tin District, Southwest China, Using GEOFRAC

Fracture is an important factor controlling mineralization and ore distribution in metallic deposits. Fracture also affects mechanical stability of drifts. A plausible three-dimensional discrete fracture network (DFN) was constructed using GEOFRAC, a geostatistical method of conditioning directions (strikes and dips) and locations of sample fractures. The northern part of the Gejiu tin district, southwest China, was selected, using 10,212 fractures sampled on drift roofs at five different levels and along a 5463 m length. Key parameters of this DFN are fracture density, direction, and the connective condition of disks that form a fracture plane. The simulated fractures were verified by good correspondence of their directions with those of the sample fractures; furthermore, continuous fractures longer than 1 km matched the faults observed in outcrops. The most notable result was clarification of fracture control by the orebody shape: Simulated gentle fractures (30° dips or less) formed layered orebodies and created boundaries between each layer, while steep fractures (60° dips or more) displaced the layers by acting as faults. Such fault-type fractures and the densely fractured portions require caution with respect to mining safety. It was concluded that GEOFRAC is useful for three-dimensional DFN modeling in underground mines.

58 GEOSCIENCES↗
DOE OSTI2014

Fifteen-Year Global Time Series of Satellite-Derived Fine Particulate Matter

Ambient fine particulate matter (PM2.5) is a leading environmental risk factor for premature mortality. We use aerosol optical depth (AOD) retrieved from two satellite instruments, MISR and SeaWiFS, to produce a unified 15-year global time series (1998-2012) of ground-level PM2.5 concentration at a resolution of 1 degrees x 1 degrees. The GEOS-Chem chemical transport model (CTM) is used to relate each individual AOD retrieval to ground-level PM2.5. Four broad areas showing significant, spatially coherent, annual trends are examined in detail: the Eastern U.S. (-0.39 +/- 0.10 mu g m(-3) yr(-1)), the Arabian Peninsula (0.81 +/- 0.21 mu g m(-3) yr(-1)), South Asia (0.93 +/- 0.22 mu g m(-3) yr(-1)) and East Asia (0.79 +/- 0.27 mu g m(-3) yr(-1)). Over the period of dense in situ observation (1999-2012), the linear tendency for the Eastern U.S. (-0.37 +/- 0.13 mu g m(-3) yr(-1)) agrees well with that from in situ measurements (-0.38 +/- 0.06 mu g m(-3) yr(-1)). A GEOS-Chem simulation reveals that secondary inorganic aerosols largely explain the observed PM2.5 trend over the Eastern U.S., South Asia, and East Asia, while mineral dust largely explains the observed trend over the Arabian Peninsula.

Boys, B. L.↗
DOE OSTI2023

A fine pore-preserved deep neural network for porosity analytics of a high burnup U-10Zr metallic fuel

Abstract U-10 wt.% Zr (U-10Zr) metallic fuel is the leading candidate for next-generation sodium-cooled fast reactors. Porosity is one of the most important factors that impacts the performance of U-10Zr metallic fuel. The pores generated by the fission gas accumulation can lead to changes in thermal conductivity, fuel swelling, Fuel-Cladding Chemical Interaction (FCCI) and Fuel-Cladding Mechanical Interaction (FCMI). Therefore, it is crucial to accurately segment and analyze porosity to understand the U-10Zr fuel system to design future fast reactors. To address the above issues, we introduce a workflow to process and analyze multi-source Scanning Electron Microscope (SEM) image data. Moreover, an encoder-decoder-based, deep fully convolutional network is proposed to segment pores accurately by integrating the residual unit and the densely-connected units. Two SEM 250 × field of view image datasets with different formats are utilized to evaluate the new proposed model’s performance. Sufficient comparison results demonstrate that our method quantitatively outperforms two popular deep fully convolutional networks. Furthermore, we conducted experiments on the third SEM 2500 × field of view image dataset, and the transfer learning results show the potential capability to transfer the knowledge from low-magnification images to high-magnification images. Finally, we use a pre-trained network to predict the pores of SEM images in the whole cross-sectional image and obtain quantitative porosity analysis. Our findings will guide the SEM microscopy data collection efficiently, provide a mechanistic understanding of the U-10Zr fuel system and bridge the gap between advanced characterization to fuel system design.

36 MATERIALS SCIENCE↗