Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “sampling variability”

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 1,639 records · Page 91
All sourcesNASA NTRSDOE OSTI
DOE OSTI2015

Bioaccessibility of Cd and Pb in tailings from a zinc smelting in Brazil: implications for human health

Soils and wastes enriched with heavy metals may present significant ecological and human health risks. A considerable number of mining/smelting areas exist in Brazil, where high levels of metals have been found, as well as in their surrounding soils and sediments. However, studies of bioaccessibility of metals in soils/tailings from these areas are scarce, despite their potential informational contribution concerning exposure risks for residents near these areas. This study evaluated tailing samples collected from four sites of a smelting area aiming to: (i) evaluate the presence of metals of potential concern; (ii) investigate Cd and Pb bioaccessibility; and, (iii) determine the desorption kinetics of Cd and Pb. Five composite samples were collected from each site, at two depths, from a zinc smelting area located near Três Marias city - MG, Brazil. Availability of Cd and Pb was measured using a bioaccessibility test and a desorption experiment. High concentrations of total Cd and Pb and great variability were found in the tailings (up to 1743 mg Cd kg -1 and 8675 mg Pb kg -1 ), indicating the importance of adequate planning for their final disposal, in order to avoid contamination in the surrounding environment. Cadmium and Pb bioaccessibility percentages in the intestinal phase were less than 47 and 4%, respectively, which represents significant fractions not available for absorption in the intestinal tract. However, this material has to be monitored since its bioaccessibility may increase with eventual physicochemical changes, releasing Cd and Pb. Desorption kinetics experiments revealed that the Pb in the samples remained in less labile fractions whereas Cd was found in more labile fractions, which is in accordance with the bioaccessibility result.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2018

Giant Modal Gain Coefficients in Colloidal II–VI Nanoplatelets

Modal gain coefficient is a key figure of merit for a laser material. Previously, net modal gain coefficients larger than a few thousand cm –1 were achieved in II–VI and III–V semiconductor gain media, but this required operation at cryogenic temperatures. In this work, using pump-fluence-dependent variable-stripe-length measurements, we show that colloidal CdSe nanoplatelets enable giant modal gain coefficients at room temperature up to 6600 cm –1 under pulsed optical excitation. Furthermore, we show that exceptional gain performance is common to the family of CdSe nanoplatelets, as shown by examining samples having different vertical thicknesses and lateral areas. Altogether, colloidal II–VI nanoplatelets with superior optical gain properties are promising for a broad range of applications, including high-speed light amplification and loss compensation in plasmonic photonic circuits.

42 ENGINEERING↗
DOE OSTI2019

Online Static Load Model Estimation in Distribution Systems: Preprint

The paper focuses on parameter identification for time varying voltage dependent polynomial load models, namely the constant impedance (Z), constant current (I), and constant power (P), in the distribution systems. Typically, parameter identification has been done for aggregated loads at substations for the transmission planning studies. Given the higher voltage volatility faced by active distribution systems, this paper takes the load model estimation to the distribution networks. The paper assumes the presence of measurements at the distribution nodes from various sources, each at varying temporal resolutions. Least squares estimation (LSE) method with bounded variables is used to convert the measurements into ZIP parameter estimations for loads at various locations of the feeder. The results and discussions focus on some of the estimation issues faced in a distribution system due to varying voltage sensitivities of loads along a feeder, measurement resolution, impact of window size and sampling rate of measurement data, and the use of dynamic simulations to create synthetic measurement data in the absence of real ones.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Identifications of RR Lyrae Stars and Quasars from the Simulated Data of Mephisto-W Survey

We have investigated the feasibilities and accuracies of the identifications of RR Lyrae stars and quasars from the simulated data of the Multi-channel Photometric Survey Telescope (Mephisto) W Survey. Based on the variable sources light curve libraries from the Sloan Digital Sky Survey (SDSS) Stripe 82 data and the observation history simulation from the Mephisto-W Survey Scheduler, we have simulated the uvgriz multi-band light curves of RR Lyrae stars, quasars and other variable sources for the first-year observation of Mephisto W Survey. We have applied the ensemble machine learning algorithm Random Forest Classifier (RFC) to identify RR Lyrae stars and quasars, respectively. We build training and test samples and extract ~150 features from the simulated light curves and train two RFCs respectively for the RR Lyrae star and quasar classification. We find that, our RFCs are able to select the RR Lyrae stars and quasars with remarkably high precision and completeness, with purity = 95.4% and completeness = 96.9% for the RR Lyrae RFC and purity = 91.4% and completeness = 90.2% for the quasar RFC. In conclusion, we have also derived relative importances of the extracted features utilized to classify RR Lyrae stars and quasars.

(galaxies:) quasars: general↗
DOE OSTI2016

Tungsten Incorporation into Gallium Oxide: Crystal Structure, Surface and Interface Chemistry, Thermal Stability and Interdiffusion

Tungsten (W) incorporated gallium oxide (Ga2O3) (GWO) thin films were deposited by radio-frequency magnetron co-sputtering of W-metal and Ga2O3-ceramic targets. Films were produced by varying sputtering power applied to the W-target in order to achieve variable W-content (0-12 at%) into Ga2O3 while substrate temperature was kept constant at 500 °C. Chemical composition, chemical valence states, microstructure and crystal structure of as-deposited and annealed GWO films were evaluated as a function of W-content. The structural and chemical analyses indicate that the samples deposited without any W-incorporation are stoichiometric, nanocrystalline Ga2O3 films, which crystallize in β-phase monoclinic structure. While GWO films also crystallize in monoclinic β-Ga2O3 phase, W-incorporation induces surface amorphization as revealed by structural studies. The chemical valence state of Ga ions probed by X-ray photoelectron spectroscopic (XPS) analyses is characterized by the highest oxidation state i.e., Ga3+. No changes in Ga chemical state are noted for variable W-incorporation in the range of 0-12 at%. Rutherford backscattering spectrometry (RBS) analyses indicate the uniform distribution of W-content in the GWO films. However, XPS analyses indicate the formation of mixed valence states for W ions, which may be responsible for surface amorphization in GWO films. GWO films were stable up to 900 oC, at which point thermally induced secondary phase (W-oxide) formation was observed. A transition to mesoporous structure coupled with W interdiffusion occurs due to thermal annealing as derived from the chemical analyses at the GWO films’ surface as well as depth-profiling towards the GWO-Si interface. A model has been formulated to account for the mechanism of W-incorporation, thermal stability and interdiffusion via pore formation in GWO films.

Chemical composition↗
DOE OSTI2014

An integrated top-down and bottom-up proteomic approach to characterize the antigen binding fragment of antibodies

We have previously shown that different individuals exposed to the same antigen produce antibodies with identical mutations in their complementarity determining regions (CDR), suggesting that CDR tryptic peptides can serve as biomarkers for disease diagnosis and prognosis. Complete Fabs derived from disease specific antibodies have even higher potential; they could potentially be used for disease treatment and are required to identify the antigens towards which the antibodies are directed. However, complete Fab sequence characterization via LC-MS analysis of tryptic peptides (i.e. bottom-up) has proven to be impractical for mixtures of antibodies. To tackle this challenge, we have developed an integrated bottom-up and top-down MS approach, employing 2D chromatography coupled with Fourier transform mass spectrometry (FTMS), and applied this approach for full characterization of the variable parts of two pharmaceutical monoclonal antibodies with sensitivity comparable to the bottom-up standard. These efforts represent an essential step towards the identification of disease specific antibodies in patient samples with potentially significant clinical impact.

Antibodies↗
DOE OSTI2018

Micromechanical processes in consolidated granular salt

Here, granular salt is likely to be used as backfill material and a seal system component within geologic salt formations serving as a repository for long-term isolation of nuclear waste. Pressure from closure of the surrounding salt formation will promote consolidation of granular salt, eventually resulting in properties comparable to native salt. Understanding dependence of consolidation processes on stress state, moisture availability, temperature, and time is important for demonstrating sealing functions and long-term repository performance. This study characterizes laboratory-consolidated granular salt by means of microstructural observations. Granular salt material from mining operations was obtained from the bedded Salado Formation hosting the Waste Isolation Pilot Plant and the Avery Island salt dome. Laboratory test conditions included hydrostatic consolidation of jacketed granular salt with varying conditions of confining isochoric stress to 38 MPa, temperature to 250 °C, moisture additions of 1% by weight, time duration, and vented and non-vented states. Resultant porosities ranged between 1% and 22%. Optical and scanning electron microscopic techniques were used to ascertain consolidation mechanisms. From these investigations, samples with 1% added moisture or unvented during consolidation, exhibit clear pressure solution processes with tightly cohered grain boundaries and occluded fluid pores. Samples with only natural moisture content consolidated by a combination of brittle, cataclastic, and crystal plastic deformation. Recrystallization at 250 °C irrespective of moisture conditions was also observed. The range and variability of conditions applied in this study, combined with the techniques used to display microstructural features, are unique, and provide insight into an important area of governing deformation mechanism(s) occurring within salt repository applications.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2018

Chemical variability in mineralized veins observed by ChemCam on the lower slopes of Mount Sharp in Gale crater, Mars

ChemCam has observed a wide range of diagenetic features along the Curiosity rover traverse including pervasive Ca-sulfate veins. Observations by multiple instruments on Curiosity indicate that these veins are hydrated, formed during diagenetic fluid event(s). In this study, we delve into the chemical variability in these Ca-sulfate bearing veins and have identified two subsets in the Murray formation with enrichments in Fe and Fe + Mg. These chemical trends do not reflect a sampling mixture with the surrounding host rock but likely indicates the presence of authigenic phases formed during the emplacement of these veins. Based on passive reflectance spectral analysis and correlation with other elements, Fe 3+ oxides and/or sulfates are proposed to account for the Fe-rich observations in the vicinity of the Naukluft Plateau whereas the Fe + Mg trend is also observed in adjacent dark-toned features with elevated Mn and P near the Old Soaker outcrop. In conclusion, the specific localization of these observations in the Gale stratigraphy implies changing pH and redox conditions in the groundwater at the time of formation of these veins, from oxidizing and likely more acidic near the Naukluft Plateau to more reducing conditions in the upper part of the Murray formation.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2018

GaN decomposition after high-energy ball milling and its transformation to BN hollow spheres in annealing samples doped with B

Highlights: • B- and Si-doped GaN semiconductor are processed by high-energy ball milling. • TEM observations indicate GaN structural decomposition. • DSC-TGA analysis confirms the GaN temperatures decomposition. • GaN decomposition is more influenced by B element. • BN hollow particles is formed after annealing treatment of the B-doped GaN. - Abstract: In this work, GaN powders doped with B and Si were treated by high-energy ball milling, and subsequently, annealing at different temperatures. Transmission electron microscopy (TEM) observations showed the in-situ GaN decomposition in the undoped GaN specimen milled for 5 h. TEM also revealed that the chemical decomposition of GaN was more affected in samples doped with B. Thermo-gravimetric analysis indicates that an undoped GaN specimen decomposes at 1120 °C, Si-doped GaN samples disintegrate at 1150 °C, and finally, B-doped GaN samples decompose at the lower temperature of 940 °C. TEM and SEM results show that the annealing treatment of the as-milled B-doped GaN samples at 1150 °C for 15 min leads to the GaN decomposition and BN hollow nanostructures formation under an N{sub 2} atmosphere. The synthesized product consists of mesoporous powders with variable particle sizes from 500 nm to 1 μm having variable cavities in the range from 60 nm to 800 nm.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Characterizing uncertainties in ground “truth” of precipitation over complex terrain through high-resolution numerical modeling

Ground observation of precipitation over complex terrain is subject to large uncertainties due to inadequate sampling. This study explores a method that combines limited gauge data and a high-resolution numerical simulation to quantify the precipitation uncertainties in central Himalaya. Specifically, the Coefficient of spatial Variability (CV) of precipitation and the minimum Number of Required Stations (NRS) to obtain areal-mean precipitation ground truth values within a 0.25-degree area are investigated using fine-scale meteorological simulation at 1.5 km grid spacing. Evaluation over a densely-gauged area demonstrates comparable CV and NRS values between station observations and simulations. The simulated CV and NRS values in a larger area show a strong and positive dependence on each other and an expected positive (negative) correlation with topographic complexity (temporal scale). The proposed method sheds lights on evaluating precipitation products and holds promise for informing the layout of rain gauge networks in complex terrain.

Ouyang, Lin↗
DOE OSTI2014

Invited Article: An integrated mid-infrared, far-infrared, and terahertz optical Hall effect instrument

We report on the development of the first integrated mid-infrared, far-infrared, and terahertz optical Hall effect instrument, covering an ultra wide spectral range from 3 cm{sup −1} to 7000 cm{sup −1} (0.1–210 THz or 0.4–870 meV). The instrument comprises four sub-systems, where the magneto-cryostat-transfer sub-system enables the usage of the magneto-cryostat sub-system with the mid-infrared ellipsometer sub-system, and the far-infrared/terahertz ellipsometer sub-system. Both ellipsometer sub-systems can be used as variable angle-of-incidence spectroscopic ellipsometers in reflection or transmission mode, and are equipped with multiple light sources and detectors. The ellipsometer sub-systems are operated in polarizer-sample-rotating-analyzer configuration granting access to the upper left 3 × 3 block of the normalized 4 × 4 Mueller matrix. The closed cycle magneto-cryostat sub-system provides sample temperatures between room temperature and 1.4 K and magnetic fields up to 8 T, enabling the detection of transverse and longitudinal magnetic field-induced birefringence. We discuss theoretical background and practical realization of the integrated mid-infrared, far-infrared, and terahertz optical Hall effect instrument, as well as acquisition of optical Hall effect data and the corresponding model analysis procedures. Exemplarily, epitaxial graphene grown on 6H-SiC, a tellurium doped bulk GaAs sample and an AlGaN/GaN high electron mobility transistor structure are investigated. The selected experimental datasets display the full spectral, magnetic field and temperature range of the instrument and demonstrate data analysis strategies. Effects from free charge carriers in two dimensional confinement and in a volume material, as well as quantum mechanical effects (inter-Landau-level transitions) are observed and discussed exemplarily.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2015

The identification of extreme asymptotic giant branch stars and red supergiants in M33 with 24 μm variability

We present the first detection of 24 μm variability in 24 sources in the Local Group galaxy M33. These results are based on 4 epochs of Multiband Imaging Photometer for Spitzer observations, which are irregularly spaced over ∼750 days. We find that these sources are constrained exclusively to the Holmberg radius of the galaxy, which increases their chances of being members of M33. We have constructed spectral energy distributions (SEDs) ranging from the optical to the submillimeter to investigate the nature of these objects. We find that 23 of our objects are most likely heavily self-obscured, evolved stars, while the remaining source is the Giant H ii region, NGC 604. We believe that the observed variability is the intrinsic variability of the central star reprocessed through their circumstellar dust shells. Radiative transfer modeling was carried out to determine their likely chemical composition, luminosity, and dust production rate (DPR). As a sample, our modeling has determined an average luminosity of (3.8±0.9)×10{sup 4} L{sub ⊙} and a total DPR of (2.3±0.1)×10{sup −5} M{sub ⊙} yr{sup −1}. Most of the sources, given the high DPRs and short wavelength obscuration, are likely extreme asymptotic giant branch (XAGB) stars. Five of the sources are found to have luminosities above the classical AGB limit (M{sub bol} <−7.1 mag, L > 54,000 L{sub ⊙}), which classifies them as probable red supergiants (RSGs). Almost all of the sources are classified as oxygen-rich. As also seen in the LMC, a significant fraction of the dust in M33 is produced by a handful of XAGB and RSG stars.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2011

THE RELATION OF OPTICAL/ULTRAVIOLET AND X-RAY EMISSION IN LOW-LUMINOSITY ACTIVE GALACTIC NUCLEI

We study the relation of optical/UV and X-ray emission in the low luminosity active galactic nuclei (LLAGNs), using a sample of 49 sources including 28 local Seyfert galaxies and 21 low-ionization nuclear emission-line regions with optical/UV spectral luminosity at wavelength {lambda} = 2500 A, 23.0 {<=} log L{sub {nu}(2500A)}(erg s{sup -1} Hz{sup -1}) {<=} 27.7, and X-ray spectral luminosity at 2 keV, 20.5 {<=} log L{sub {nu}(2keV)} {<=} 25.3. Strong correlations are found between the X-ray luminosity and the optical/UV-to-X-ray index, {alpha}{sub ox}, with the optical/UV luminosity, with slopes very similar to the findings for the luminous AGNs in previous works. The correlation between {alpha}{sub ox} and L{sub {nu}(2keV)} is very weak, as is that found for the luminous AGNs in the majority of previous similar works. We also study the relation between {alpha}{sub ox} and the Eddington ratio L{sub bol}/L{sub Edd} for our sample and find a significant anti-correlation for the sources with L{sub bol}/L{sub Edd} {approx}< 10{sup -3}, which is opposite to the correlation between the two variables for the luminous AGNs. Using the advection-dominated accretion flow (ADAF) model, we roughly reproduce this anti-correlationship for the two variables for the LLAGNs. This result strongly supports the ADAF as a candidate for the accretion mode in LLAGNs.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2014

Investigating the variability of active galactic nuclei using combined multi-quarter Kepler data

We used photometry from the Kepler satellite to characterize the variability of four radio-loud active galactic nuclei (AGNs) on timescales from years to minutes. The Kepler satellite produced nearly continuous high precision data sets which provided better temporal coverage than possible with ground based observations. We have now accumulated 11 quarters of data, eight of which were reported in our previous paper. In addition to constructing power spectral densities (PSDs) and characterizing the variability of the last three quarters, we have linked together the individual quarters using a multiplicative scaling process, providing data sets spanning ∼2.8 yr with >98% coverage at a 30 minute sampling rate. We compute PSDs on these connected data sets that yield power law slopes at low frequencies in the approximate range of –1.5 to –2.0, with white noise seen at higher frequencies. These PSDs are similar to those of both the individual quarters and to those of ground-based optical observations of other AGNs. We also have explored a PSD binning method intended to reduce a bias toward shallow slope fits by evenly distributing the points within the PSDs. This tends to steepen the computed PSD slopes, especially when the low frequencies are relatively poorly fit. We detected flares lasting several days in which the brightness increased by ∼15%-20% in one object, as well a smaller flare in another. Two AGNs showed only small, ∼1%-2%, fluctuations in brightness.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Evidence for multi-decadal fuel buildup in a large California wildfire from smoke radiocarbon measurements

In recent decades, there has been a significant increase in annual area burned in California's Sierra Nevada mountains. This rise in fire activity has prompted the need to understand how historical forest management practices affect fuel composition and emissions. Here we examined the total carbon (TC) concentration and radiocarbon abundance (Δ 14 C) of particulate matter (PM) emitted by the KNP Complex Fire, which occurred during California's 2021 wildfire season and affected several groves of giant sequoia trees in the southern Sierra Nevada. During a 26 h sampling period, we measured concentrations of fine airborne PM (PM 2.5 ), as well as dry air mole fractions of carbon monoxide (CO) and methane (CH 4 ), using a ground-based mobile laboratory. We also collected filter samples of PM 2.5 for analysis of TC concentration and Δ 14 C. High correlation among PM 2.5 , CO, and CH 4 time series confirmed that our PM 2.5 measurements captured variability in wildfire emissions. Using a Keeling plot approach, we determined that the mean Δ 14 C of PM 2.5 was 111.6 ± 7.7‰ (n = 12), which was considerably enriched relative to atmospheric carbon dioxide in the northern hemisphere in 2021 (-3.2 ± 1.4‰). Combining these Δ 14 C data with a steady-state one-box ecosystem model, we estimated that the mean age of fuels combusted in the KNP Complex Fire was 40 years, with a range of 29–57 years. These results provide evidence for emissions originating from woody biomass, larger-diameter fine fuels, and coarse woody debris that have accumulated over multiple decades. This is consistent with independent field observations that indicate high fire intensity contributed to widespread giant sequoia mortality. With the expanded use of prescribed fires planned over the next decade in California to mitigate wildfire impacts, our measurement approach has the potential to provide regionally-integrated estimates of the effectiveness of fuel treatment programs.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2015

SOFIA OBSERVATIONS OF SN 2010jl: ANOTHER NON-DETECTION OF THE 9.7 μm SILICATE DUST FEATURE

We present photometric observations from the Stratospheric Observatory for Infrared Astronomy (SOFIA) at 11.1 μm of the Type IIn supernova (SN IIn) 2010jl. The SN is undetected by SOFIA, but the upper limits obtained, combined with new and archival detections from Spitzer at 3.6 and 4.5 μm, allow us to characterize the composition of the dust present. Dust in other SN IIn has been shown in previous works to reside in a circumstellar shell of material ejected by the progenitor system in the few millenia prior to explosion. Our model fits show that the dust in the system shows no evidence for the strong, ubiquitous 9.7 μm feature from silicate dust, suggesting the presence of carbonaceous grains. The observations are best fit with 0.01–0.05 M{sub ⊙} of carbonaceous dust radiating at a temperature of ∼550–620 K. The dust composition may reveal clues concerning the nature of the progenitor system, which remains ambiguous for this subclass. Most of the single star progenitor systems proposed for SNe IIn, such as luminous blue variables, red supergiants, yellow hypergiants, and B[e] stars, all clearly show silicate dust in their pre-SN outflows. However, this post-SN result is consistent with the small sample of SNe IIn with mid-IR observations, none of which show signs of emission from silicate dust in their IR spectra.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

TESS cycle 1 observations of roAp stars with 2-min cadence data

ABSTRACT We present the results of a systematic search for new rapidly oscillating Ap (roAp) stars using the 2-min cadence data collected by the Transiting Exoplanet Survey Satellite (TESS) during its Cycle 1 observations. We identify 12 new roAp stars. Amongst these stars we discover the roAp star with the longest pulsation period, another with the shortest rotation period, and six with multiperiodic variability. In addition to these new roAp stars, we present an analysis of 44 known roAp stars observed by TESS during Cycle 1, providing the first high-precision and homogeneous sample of a significant fraction of the known roAp stars. The TESS observations have shown that almost 60 per cent (33) of our sample of stars are multiperiodic, providing excellent cases to test models of roAp pulsations, and from which the most rewarding asteroseismic results can be gleaned. We report four cases of the occurrence of rotationally split frequency multiplets that imply different mode geometries for the same degree modes in the same star. This provides a conundrum in applying the oblique pulsator model to the roAp stars. Finally, we report the discovery of non-linear mode interactions in α Cir (TIC 402546736, HD 128898) around the harmonic of the principal mode – this is only the second case of such a phenomenon.

Holdsworth, D. L.↗
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↗