SEARCH · Engineering Papers

Results for “linear scaling”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,657 records · Page 92

All sources NASA NTRS DOE OSTI

DOE OSTI2025

Mapping Spiking Neural Networks to Heterogeneous Crossbar Architectures using Integer Linear Programming

Advances in novel hardware devices and architectures allow Spiking Neural Network (SNN) evaluation using ultra-low power, mixed-signal, memristor crossbar arrays. As individual network sizes quickly scale beyond the dimensional capabilities of single crossbars, networks must be mapped onto multiple crossbars. Crossbar sizes within modern Memristor Crossbar Architectures (MCAs) are determined predominately not by device technology but by network topology; more, smaller crossbars consume less area thanks to the high structural sparsity found in larger, brain-inspired SNNs. Motivated by continuing increases in SNN sparsity due to improvements in training methods, we propose utilizing heterogeneous crossbar sizes to further reduce area consumption. This approach was previously unachievable as prior compiler studies only explored solutions targeting homogeneous MCAs. Our work improves on the state-of-the-art by providing Integer Linear Programming (ILP) formulations supporting arbitrarily heterogeneous architectures. By modeling axonal interactions between neurons, our methods produce better mappings while removing inhibitive a priori knowledge requirements. We first show a 16.7-27.6% reduction in area consumption for square-crossbar homogeneous architectures. Then, we demonstrate 66.9-72.7% further reduction when using a reasonable configuration of heterogeneous crossbar dimensions. Next, we present a new optimization formulation capable of minimizing the number of inter-crossbar routes. When applied to solutions already near-optimal in area, an 11.9-26.4% routing reduction is observed without impacting area consumption. Finally, we present a profile-guided optimization capable of minimizing the number of runtime spikes between crossbars. Compared to the best-area-then-route optimized solutions, we observe a further 0.5-14.8% inter-crossbar spike reduction while requiring 1–3 orders of magnitude less solver time.

Pohl, Devin [ORNL] (ORCID:0009000040149027)↗

DOE OSTI2023

Scaled ILU Smoothers for Navier-Stokes Pressure Projection

Incomplete LU (ILU) smoothers are effective in the algebraic multigrid (AMG) V-cycle for reducing high-frequency components of the error. However, the requisite direct triangular solves are comparatively slow on GPUs. Previous work has demonstrated the advantages of Jacobi iteration as an alternative to direct solution of these systems. Depending on the threshold and fill-level parameters chosen, the factors can be highly nonnormal and Jacobi is unlikely to converge in a low number of iterations. We demonstrate that row scaling can reduce the departure from normality, allowing us to replace the inherently sequential solve with a rapidly converging Richardson iteration. There are several advantages beyond the lower compute time. Scaling is performed locally for a diagonal block of the global matrix because it is applied directly to the factor. Further, an ILUT Schur complement smoother maintains a constant GMRES iteration count as the number of MPI ranks increases, and thus parallel strong-scaling is improved. Our algorithms have been incorporated into hypre, and we demonstrate improved time to solution for linear systems arising in the Nalu-Wind and PeleLM pressure solvers. For large problem sizes, GMRES+AMG executes at least five times faster when using iterative triangular solves compared with direct solves on massively parallel GPUs.

algebraic multigrid↗

DOE OSTI2023

Urbanization effects on the spatial patterns of spring vegetation phenology depend on the climatic background

Cities have been considered ideal surrogates for evaluating ecological responses to climate warming. Although research has revealed that the urban heat island effect is not the only determinant that drives the rural-urban difference in spring phenology, limited attempts further explore the effect of complex interactions between urbanization and climates on the start of the season (SOS). Here, we employed the percentage of impervious surface area as an indicator of urbanization levels, incorporating the urban heat island (UHI) effect and climates to decipher the multiple effects on the temporal shift in SOS. Our results suggest that urbanization, the UHI effect, and climates jointly drive the phenological timing in cities. Further, the SOS is highly linearly correlated with urbanization levels and UHI effects (p<0.001) in the zones with warm and humid climates, whereas relatively warmer, cooler, or arid climates break down the linearity. Geographically, the SOS is north-south symmetrical for the rural-urban ecosystem at the national scale, which initiates at the mid-latitudes (around 30 °N, 67.58 DOY ± 2.23 days) first and then towards the low (around 20 °N, 105.25 DOY ± 1.31 days) and high-latitudes (45 °N, 123.01 DOY ± 1.45 days). Chilling accumulation and spring temperature jointly contribute to this phenomenon. As the UHI effect is similar to the projected global warming in the future, understanding the phenological responses to urbanization in various climates is insightful for evaluating the impact of future warming on plants' phenological behavior in the global ecosystem.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2024

A fully implicit, asymptotic-preserving, semi-Lagrangian algorithm for the time dependent anisotropic heat transport equation

In this paper, we extend the operator-split asymptotic-preserving, semi-Lagrangian algorithm for time dependent anisotropic heat transport equation proposed in Chacón et al. (2014) [18] to use a fully implicit time integration with backward differentiation formulas. The proposed implicit method can deal with arbitrary heat-transport anisotropy ratios $\mathcal{X}$∥ /$ \mathcal{X}$⟂ $\ggg$ 1 (with $\mathcal{X}$∥, $ \mathcal{X}$⟂ the parallel and perpendicular heat diffusivities, respectively) in complicated magnetic field topologies in an accurate and efficient manner. Further, the implicit algorithm is second-order accurate temporally and demonstrates an accurate treatment at boundary layers (e.g., island separatrices), which was not ensured by the operator-split implementation. The condition number of the resulting algebraic system is independent of the anisotropy ratio, and is inverted with preconditioned GMRES. We propose a simple preconditioner that renders the finite-dimensional linear operator compact, resulting in mesh-independent convergence rates for topologically simple magnetic fields, and convergence rates scaling as ~ (NΔt) 1/4 (with N the total mesh size and Δt the timestep) in topologically complex magnetic-field configurations. We demonstrate the accuracy and performance of the approach with test problems of varying complexity, including an analytically tractable boundary-layer problem in a straight magnetic field, and a topologically complex magnetic field featuring magnetic islands with extreme anisotropy ratios $\mathcal{X}$∥ /$ \mathcal{X}$⟂ = 10 10 ) .

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

Emergence of local scaling relations in adsorption energies on high-entropy alloys

Alloying has been proposed to circumvent scaling relations between the adsorption energies thus allowing for the complete optimization of multistep reactions. Herein the fidelity of scaling rules on high-entropy alloy (HEA) surfaces is assessed focusing on hydrogen-containing molecules, *AH x for A = C and N (x = 0, 1, 2, 3), A = S (x = 0, 1, 2) and A = O (x = 0, 1). Using an adsorbate- and site-specific deep learning model to rapidly compute the adsorption energies on CoMoFeNiCu HEA surfaces, the energies of *AH x and *A are shown to be linearly correlated if *A and *AH x have identical adsorption site symmetry. However, a local linear dependence emerges between the configuration-averaged adsorption energies irrespective of the site symmetry. Although these correlations represent a weaker form of the scaling relationships, they are sufficient to prohibit the optimization of multistep reactions. The underpinning of this behavior is twofold (1) the nearsightedness principle and (2) the narrow distribution of the adsorption energies around the mean-field value. While the nearsightedness is general for all electronic systems, the second criterion applies in HEAs with relatively strong reactive elements. The present findings strongly suggest that alloys may not generally enable the breaking of scaling relationships.

36 MATERIALS SCIENCE↗

DOE OSTI2020

Very massive tracers and higher derivative biases

Most of the upcoming cosmological information will come from analyzing the clustering of the Large Scale Structures (LSS) of the universe through LSS or CMB observations. It is therefore essential to be able to understand their behavior with exquisite precision. The Effective Field Theory of Large Scale Structures (EFTofLSS) provides a consistent framework to make predictions for LSS observables in the mildly non-linear regime. Herein we focus on biased tracers. We argue that in calculations at a given order in the dark matter perturbations, highly biased tracers will underperform because of their larger higher derivative biases. A natural prediction of the EFTofLSS is therefore that by simply adding higher derivative biases, highly massive tracers should perform comparably well. We implement this prediction for the halo-halo and the halo-matter power spectra at one loop, and the halo-halo-halo, halo-halo-matter, and halo-matter-matter bispectra at tree-level, and compare with simulations. We find good agreement with the prediction: at $z$ = 0, for all tracers, we are able to match the power spectra up to $k ≃$ 0.28$h$ Mpc -1 , as well as a small set of about 10 2 bispectra triangles up to $k ≃$ 0.17$h$ Mpc -1 . We also discuss the limitations of our study and some avenues to pursue to further establish these findings.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2025

Impact of ionization peak location on measured opaqueness in DIII-D H-mode plasmas

This study investigates the relationship between electron pedestal density and the location of the ionization peak on neutral penetration in DIII-D H-mode plasmas, utilizing a database of Lyman-α emission measurements. The high electron density leads to neutrals being ‘screened’ and the ionization front being pushed out into the Scrape-Off Layer (SOL). This is also referred to as the neutral opaqueness, which is heuristically expected to scale with edge plasma density and machine size. However, at lower electron pedestal density, the penetration depth of the neutrals varies, and measured opaqueness deviates from the heuristic scaling. The database reveals that at low density, when the ionization peak is located in SOL region, the linear relationship between the electron density and neutral penetration holds. However, when the peak is located inside the separatrix, the penetration of the neutrals (λ n 0 ) is much wider ~3.0–3.5 cm, breaking the heuristic opaqueness approximation. These findings provide valuable insights into fueling efficiency and plasma behavior, with implications for Fusion Pilot Plants where high pedestal densities are anticipated and where the neutral opaqueness behaves like its heuristic approximation. This analysis offers a framework to refine neutral opaqueness approximations, enhancing the predictive capability for advanced tokamak operations.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2026

Systematic multi-machine analysis of the exhaust time-dependent behavior in tokamaks

The understanding of the time-scales and associated transient behavior of fusion exhaust plasmas plays a crucial role in its dynamic modeling and its control. This work presents an overview of experimental investigations of the exhaust dynamics in TCV, MAST-U, ASDEX-Upgrade, WEST, DIII-D, and JET. From the presented experiments, a clear picture arises on properties of the exhaust dynamics across machines. Particularly, we observe that the scrape-off layer equilibrates on fast time-scales ($>$ 70 Hz) and that exhaust dynamics measured in response to gas valve modulations mostly behave smoothly and linearly, with similarities across devices, across scenarios (H-mode, L-mode), injected species, and injection locations. The measurements presented have formed the basis for systematic exhaust control on the considered devices. We now present this database for the essential validation of dynamic exhaust models for reactor design and control.

control↗

DOE OSTI2020

Probing the structure of the initial state of heavy-ion collisions with p T -dependent flow fluctuations

The connection between initial-state geometry and anisotropic flow can be quantified through a well-established mapping between p T -integrated flow harmonics and cumulants of the initial transverse energy distribution. In this paper we successfully extend this mapping to also include p T -differential flow. In doing so, we find that subleading principal components of anisotropic flow can reveal previously unobserved details of the hydrodynamic response, in both the linear and the nonlinear regimes. Most importantly, we show that they provide novel information on the small-scale structures present in the initial stage of relativistic heavy-ion collisions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2022

Searching for axionlike time-dependent cosmic birefringence with data from SPT-3G

Ultralight axionlike particles (ALPs) are compelling dark matter candidates because of their potential to resolve small-scale discrepancies between Λ CDM predictions and cosmological observations. Axion-photon coupling induces a polarization rotation in linearly polarized photons traveling through an ALP field; thus, as the local ALP dark matter field oscillates in time, distant static polarized sources will appear to oscillate with a frequency proportional to the ALP mass. We use observations of the cosmic microwave background from SPT-3G, the current receiver on the South Pole Telescope, to set upper limits on the value of the axion-photon coupling constant g φ γ over the approximate mass range 10 - 22 – 10 - 19 eV , corresponding to oscillation periods from 12 hours to 100 days. For periods between 1 and 100 days ( 4.7 × 10 - 22 eV ≤ m φ ≤ 4.7 × 10 - 20 eV ), where the limit is approximately constant, we set a median 95% C.L. upper limit on the amplitude of on-sky polarization rotation of 0.071 deg. Assuming that dark matter comprises a single ALP species with a local dark matter density of 0.3 GeV / cm 3 , this corresponds to g φ γ < 1.18 × 10 - 12 GeV - 1 × ( m φ 1.0 × 10 - 21 eV ) . These new limits represent an improvement over the previous strongest limits set using the same effect by a factor of ~ 3.8 .

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

Foreground-immune CMB lensing reconstruction with polarization

Extragalactic foregrounds are known to generate significant biases in temperature based CMB lensing reconstruction. Several techniques, which include "source hardening'" and "shear-only estimators'' have been proposed to mitigate contamination and have been shown to be very effective at reducing foreground-induced biases. Here we extend both techniques to polarization, which will be an essential component of CMB lensing reconstruction for future experiments and investigate the "large lens'' limit analytically to gain insight on the origin and scaling of foreground biases, as well as the sensitivity to their profiles. Furthermore, using simulations of polarized point sources, we estimate the expected bias to both Simons Observatory and CMB-S4 like (polarization-based) lensing reconstruction, finding that biases to the former are minuscule while those to the latter are potentially non-negligible at small scales ($L$ ~ $1000-2000$). In particular, we show that for a CMB-S4 like experiment, an optimal linear combination of point-source hardened estimators can reduce the (point-source induced) bias to the CMB lensing power spectrum by up to two orders of magnitude, at a ~ 4% noise cost relative to the global minimum variance estimator.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Characterization of sub-micrometre-sized voids in fixed human brain tissue using scanning X-ray microdiffraction

Using a 5 µm-diameter X-ray beam, we collected scanning X-ray microdiffraction in both the small-angle (SAXS) and the wide-angle (WAXS) regimes from thin sections of fixed human brain tissue from Alzheimer's subjects. The intensity of scattering in the SAXS regime of these patterns exhibits essentially no correlation with the observed intensity in the WAXS regime, indicating that the structures responsible for these two portions of the diffraction patterns, which reflect different length scales, are distinct. SAXS scattering exhibits a power-law behavior in which the log of intensity decreases linearly with the log of the scattering angle. The slope of the log–log curve is roughly proportional to the intensity in the SAXS regime and, surprisingly, inversely proportional to the intensity in the WAXS regime. We interpret these observations as being due to the presence of sub-micrometre-sized voids formed during dehydration of the fixed tissue. The SAXS intensity is due largely to scattering from these voids, while the WAXS intensity derives from the secondary structures of macromolecular material surrounding the voids. The ability to detect and map the presence of voids within thin sections of fixed tissue has the potential to provide novel information on the degradation of human brain tissue in neurodegenerative diseases.

Chemistry↗

DOE OSTI2026

Design and fabrication of ion traps for low RF power dissipation

Large surface-electrode ion traps with multiple trapping regions and junctions are a natural approach to scaling trapped ion quantum computers, supporting the connectivity and ion counts necessary for complex quantum algorithms. However, a major hurdle in this scaling is on-chip power dissipation from the applied RF voltage, which increases at a rate between linear and cubic relative to trap size, depending on whether the losses are dielectric or Ohmic. Here, we present two versions of a trap with features designed to reduce both types of RF power dissipation. The first variant contains a raised RF electrode that increases the electrode–ground distance to reduce capacitance. Different DC voltage sources are demonstrated on this trap to show that technical noise before the filter remains the dominant source of voltage noise and therefore motional heating. The second variant additionally includes a method for removing dielectric from beneath the RF electrode to further reduce dielectric losses. These traps were demonstrated at room temperature with 40 Ca + ions. In conclusion, the similar heating rates and heating rate axial frequency dependencies between 2.4 and 3.0 MHz illustrate that this dielectric modification is not detrimental to trap performance.

Sterk, J. D. [Sandia National Laboratories (SNL-NM↗

DOE OSTI2020

Scalable Asynchronous Domain Decomposition Solvers

We discuss how parallel implementations of linear iterative solvers generally alternate between phases of data exchange and phases of local computation. Increasingly large problem sizes and more heterogeneous compute architectures make load balancing and the design of low latency network interconnects that are able to satisfy the communication requirements of linear solvers very challenging tasks. In particular, global communication patterns such as inner products become increasingly limiting at scale. We explore the use of asynchronous communication based on one-sided Message Passing Interface primitives in the context of domain decomposition solvers. In particular, a scalable asynchronous two-level Schwarz method is presented. We discuss practical issues encountered in the development of a scalable solver and show experimental results obtained on a state-of-the-art supercomputer system that illustrate the benefits of asynchronous solvers in load balanced as well as load imbalanced scenarios. Using the novel method, we can observe speedups of up to four times over its classical synchronous equivalent.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Parallel Algebraic Multigrid for Fusion and Higher-Order PDEs

Multigrid methods play a key role in large-scale scientific simulation because they are among the fastest and most scalable approaches for solving the underlying sparse linear systems of equations that arise from a wide array of Partial Differential Equation (PDE) discretizations. Algebraic multigrid (AMG) is a special type of multigrid method that depends only on the description of the linear system, giving it better portability and broader applicability than geometric multigrid, as it requires no explicit knowledge of the problem geometry. Even though these methods are widely used today, there are still applications where further development is needed. In this report, we focus on PDEs with higher-order terms (e.g., fourth order), concentrating on a PDE that arises in tokamak edge plasma simulations (a tokamak is a machine that confines a plasma using magnetic fields and is believed to be the leading plasma confinement concept for future fusion power plants). General multigrid relaxes a linear system on coarser grids and reverses this process with interpolation, but standard AMG methods struggle with the aforementioned higher-order PDEs. We investigate cyclic coarsening and interpolation heuristics, as well as new iterative approximation methods of refining the solution at each grid to improve the existing multigrid approach. To this end, we ensure that these techniques are transferable to a parallelized setting with LLNL’s supercomputers.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2026

Advancing the quantification of aerosol-cloud interactions with the CALIPSO-CloudSat-Aqua/MODIS record

Aerosol-cloud-precipitation interactions are assessed over the non-polar ocean using more than 11 years of combined Aqua-MODIS, CALIPSO-CALIOP, and CloudSat products. The analysis first shows the benefit of incorporating vertically resolved aerosol extinction coefficient (σext) in aerosol-cloud interactions (ACI) assessments, demonstrating that: σext vertically collocated with the cloud layer () correlates best with cloud droplet number concentration (Nd), column-integrated aerosol optical depth (AOD) cannot explain the Nd variability in the extratropics, and the S-shape of the AOD-Nd relationship reported in previous studies is not replicated when using instead of AOD, with a Nd- linearity more consistent with in-situ studies over the ocean. ACI metric, estimated as the log-scale regression between CALIOP and MODIS Nd reveals that the eastern Pacific is the region with the strongest ACI, followed by the Southern Ocean. The susceptibility of clouds to changes in their liquid water path (LWP) and frequency of precipitation followed a 2-step calculation by combining the Nd- regression (ACI) with the regression between these macrophysical variables and Nd. LWP susceptibility is negative (LWP decreases with aerosol loading) and statistically significant over the eastern Pacific, eastern Atlantic, and extratropics. In contrast, vast areas of the tropical and subtropical ocean feature negligible changes in LWP with aerosol. Precipitation frequency susceptibility is negative, but the values are only significant over the coastal eastern Pacific and Atlantic. The findings suggest that previous modeling assessments relying on AOD may need to be revisited by taking advantage of the synergy between passive and active sensors.

Li, Zhujun↗

NASA NTRS1973

Prediction of free turbulent mixing using a turbulent kinetic energy method

Free turbulent mixing of two-dimensional and axisymmetric one- and two-stream flows is analyzed by a relatively simple turbulent kinetic energy method. This method incorporates a linear relationship between the turbulent shear and the turbulent kinetic energy and an algebraic relationship for the length scale appearing in the turbulent kinetic energy equation. Good results are obtained for a wide variety of flows. The technique is shown to be especially applicable to flows with heat and mass transfer, for which nonunity Prandtl and Schmidt numbers may be assumed.

Harsha, P. T.↗

NASA NTRS1977

Foster-Miller's development of dry coal feed systems

Dry coal feeder systems developed for pressurized conversion processes were carried through a laboratory scale development program. These concepts include: (1) a centrifugal solids feeder; (2) a fluidized piston feeder; (3) a linear pocket feeder; and (4) a compacted coal plug feeder. Results of laboratory model testing of all concepts are reviewed.

Harding, J. C.↗