Search NASA⌕ Search

SEARCH · Search NASA

Results for “Operational adaptability”

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 955 records · Page 53

Integrating Reanalysis and Satellite Cloud Information to Estimate Surface Downward Long-Wave Radiation

The estimation of downward long-wave radiation (DLR) at the surface is very important for the understanding of the Earth’s radiative budget with implications in surface–atmosphere exchanges, climate variability, and global warming. Theoretical radiative transfer and observationally based studies identify the crucial role of clouds in modulating the temporal and spatial variability of DLR. In this study, a new machine learning algorithm that uses multivariate adaptive regression splines (MARS) and the combination of near-surface meteorological data with satellite cloud information is proposed. The new algorithm is compared with the current operational formulation used by the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT) Satellite Application Facility on Land Surface Analysis (LSA-SAF). Both algorithms use near-surface temperature and dewpoint temperature along with total column water vapor from the latest European Centre for Medium-range Weather Forecasts (ECMWF) reanalysis ERA5 and satellite cloud information from the Meteosat Second Generation. The algorithms are trained and validated using both ECMWF-ERA5 and DLR acquired from 23 ground stations as part of the Baseline Surface Radiation Network (BSRN) and the Atmospheric Radiation Measurement (ARM) user facility. Results show that the MARS algorithm generally improves DLR estimation in comparison with other model estimates, particularly when trained with observations. When considering all the validation data, root mean square errors (RMSEs) of 18.76, 23.55, and 22.08 W·m –2 are obtained for MARS, operational LSA-SAF, and ERA5, respectively. The added value of using the satellite cloud information is accessed by comparing with estimates driven by ERA5 total cloud cover, showing an increase of 17% of the RMSE. The consistency of MARS estimate is also tested against an independent dataset of 52 ground stations (from FLUXNET2015), further supporting the good performance of the proposed model.

54 ENVIRONMENTAL SCIENCES↗

Shifts in hydropower operation to balance wind and solar will modify effects on aquatic biota

To avoid negative consequences to freshwater biota from climate change, society must complete the transition from fossil to renewable electricity sources. However, temporal patterns in hydropower generation (and flow releases that affect aquatic biota) may change with increased wind and solar penetration. We used power cost modeling to characterize current and future within-day and seasonal patterns in hydropower generation across the Eastern Interconnection in a wet and a dry year. Compared to the baseline, future hydropower generation across the grid decreased during the day and increased before dawn and after dusk. At a project level, such a pattern would suggest 'double peaking' operation (up- and down-ramping before dawn and after dusk, with lower releases midday). Variation in generation was higher in wet years than dry years, foreshadowing possible flow constraints on hydropower flexibility. At the grid scale, projected ramping rates were higher in all seasons. A review of the ecological literature suggests that these changes would shift the timing of invertebrate drift and elevate the risk of nest scouring during up-ramping and the risk of stranding or dewatering during down ramping. Thermal conditions may be moderated by increased ramping. Strategies for adapting to future shifts in the renewable portfolio range from re-regulation in reservoir cascades to providing flow refuge (structures and vegetation) below individual projects. Coordinated basin-scale operation can distribute peaking operation to maintain grid support while restricting local ramping at critical ecological times. In addition, research to design hybrid renewable systems that add battery storage is needed to understand how we can mitigate future risks to aquatic communities while promoting the use of renewable energy. This study, which is among the first to examine ecological side-effects of the shift to renewable energy in freshwater ecosystems, lays out a path toward understanding and navigating changes to flow regimes under the energy transition.

54 ENVIRONMENTAL SCIENCES↗

A multiscale restriction-smoothed basis method for high contrast porous media represented on unstructured grids

A wide variety of multiscale methods have been proposed in the literature to reduce runtime and provide better scaling for the solution of Poisson-type equations modeling flow in porous media. We present a new multiscale restricted-smoothed basis (MsRSB) method that is designed to be applicable to both rectilinear grids and unstructured grids. Like many other multiscale methods, MsRSB relies on a coarse partition of the underlying fine grid and a set of local prolongation operators (multiscale basis functions) that map unknowns associated with the fine grid cells to unknowns associated with blocks in the coarse partition. These mappings are constructed by restricted smoothing: Starting from a constant, a localized iterative scheme is applied directly to the fine-scale discretization to compute prolongation operators that are consistent with the local properties of the differential operators. The resulting method has three main advantages: First of all, both the coarse and the fine grid can have general polyhedral geometry and unstructured topology. This means that partitions and good prolongation operators can easily be constructed for complex models involving high media contrasts and unstructured cell connections introduced by faults, pinch-outs, erosion, local grid refinement, etc. In particular, the coarse partition can be adapted to geological or flow-field properties represented on cells or faces to improve accuracy. Secondly, the method is accurate and robust when compared to existing multiscale methods and does not need expensive recomputation of local basis functions to account for transient behavior: Dynamic mobility changes are incorporated by continuing to iterate a few extra steps on existing basis functions. This way, the cost of updating the prolongation operators becomes proportional to the amount of change in fluid mobility and one reduces the need for expensive, tolerance-based updates. Finally, since the MsRSB method is formulated on top of a cell-centered, conservative, finite-volume method, it is applicable to any flow model in which one can isolate a pressure equation. Herein, we only discuss single and two-phase incompressible models. Compressible flow, e.g., as modeled by the black-oil equations, is discussed in a separate paper.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Oxidation Behavior and Property Degradation of Nuclear Graphites

During its multidecade operation in the core of nuclear reactors, graphite components are subjected to aggressive and continuous exposure to a high field of ionizing and neutron irradiation, high temperature, and various types of present and postulated chemical attacks. High density, high crystallinity polygranular synthetic graphite is unique among other materials for its extraordinary capacity of resisting and adapting to the aggression inflicted by high temperature, high energy neutron bombardment and ionizing gamma radiation. But, as a carbonaceous material, even though of very high purity, graphite is reactive towards common oxidizing agents: oxygen, carbon dioxide, water. Safe operation of HTGRs relies, among other aspects, on engineered safeguard systems for efficient and continuous protection of graphite components against oxidation. Graphite oxidation behavior was, and continues to be, an important direction of theoretical and experimental research, engineering analyses, models and simulations, and design and safety regulations. The avalanche of publications, reports, experimental data, computer codes, and regulatory documents related to oxidation behavior of nuclear graphite is now accelerating to new levels, prompted by the increased interest for nuclear energy as a clean, carbon-free energy source. Even though public’s perception of nuclear energy advantages may still be influenced by the memories of past accidents of nuclear reactors from generations II and III, the community of informed scientists and engineers, regulators and statemen knows that generation IV of nuclear reactors is designed at very high safety standards, doubled by great advances of scientific knowledge and technological progress. One of routes of these recent advances is directed at better understanding of graphite oxidation behavior, its relationship with graphite manufacturing and microstructural properties, along with the effects of various environmental factors and process variables. Together, the recent progress in manufacturing, properties characterization, and modeling of intricated physical and chemical processes that concur to the oxidation behavior led to development of powerful simulation codes able to analyze various scenarios of normal operation and hypothetical off-normal events, and thus to clearly specify the allowable parameters envelopes for the designers, constructors, and operators of current and future modular HTGRs. This review begins with an introduction on manufacturing methods, structure, and properties of nuclear graphite, including basic requirements that this specialty graphite type must satisfy for nuclear use. It continues with a chapter on environmental effects on nuclear graphite, where emphasis is placed less on irradiation and much more on oxidation phenomena, their safety implications, and the basic traits of chronic and acute oxidation by air (oxygen) and water (humidity, steam). Particular attention is placed on the three graphite grades of interest for this document (IG-110, NBG-18, PCEA). A chapter on properties degradation induced by oxidation follows, with focus on density, dimensional, and mechanical properties changes. The next chapter is intended as a brief review of various approaches used for modeling of graphite oxidation behavior. It summarizes the progress of oxidation models, from the early attempts to complex computational approaches interfaced with specialized computer codes designed for nuclear reactor simulations. Last, a list is presented of knowledge gaps where more research is needed. A short summary concludes the review.

36 MATERIALS SCIENCE↗

Double-Shell Tank Primary Bottom Inspection Development - 19067

Double-Shell Tank (DST) integrity at Hanford is maintained with a variety of activities such as ultrasonic and visual inspections, chemistry controls, corrosion monitoring and structural analyses methods [1]. A pressing need for the Double-Shell Tank Integrity Program is the ability to inspect the bottom of the tanks. Though the DSTs' annulus spaces (the space between the internal and external tank walls) have been inspected visually since 1992 and with Ultrasonic Testing (UT) since 1996, WRPS has been unable to conduct inspections of the tank bottoms due to limitations of available technology, omitting about 40 percent of the tank surface below the maximum fill level from routine inspection. Visual inspections of the tank annulus have been conducted up to five times in some DSTs and the annulus space for all of the DSTs has been inspected twice with UT. The annulus region includes the walls of the primary tank and secondary liner along with accessible portions of the secondary liner bottom. Due to the lack of under-tank inspection technology, degradation of the primary tank bottom in Tank AY-102 was undetected until the tank leaked into its annulus space, confirmed in 2012 [2]. Since that time, WRPS has been working with the nondestructive examination community and the Pacific Northwest National Laboratory to evaluate potential technologies to conduct these challenging remote inspections. Recent work has included the development, testing, and deployment of several remote visual inspection technologies to observe the condition of the primary tank bottom of the DSTs by leveraging air distribution slots in the refractory pad on which it rests. Robotic solutions to conduct this inspection were developed with a detailed understanding of the physical configuration and operational logistics. Understanding and subsequent communication of these details to vendors within the robotics and nondestructive examination community was a critical component to the successful production of these new tools. In total, two remote visual inspection tools were produced by two vendors. The first tool was designed to leverage existing ultrasonic testing magnetic wall crawler technology in use within the double-shell tank annuli at Hanford. This resulted in a cost savings, immediate familiarization with the equipment, and a more streamlined shop to field experience. The second tool was developed from the ground up, and represents a more tailored and purpose-built solution to the problem. While costlier and more complex, it is also more versatile and capable of future growth or adaptation. Initial design of both tools began in early FY17 with fabrication and testing completed in early FY18. The tools were designed to be deployed alongside existing inspection operations and expand coverage. Development of visual inspection tools to evaluate the primary tank bottom region of the double-shell tanks was planned as the first stage of inspection. Continued development of volumetric sensors and incorporation of these inspection tools into Hanford's ongoing Double-Shell Tank Integrity Program are the next steps, continuously seeking to expand understanding of asset integrity and remaining useful life. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Online Monitoring To Enable Improved Diagnostics, Prognostics and Maintenance

Only time will tell what the implications of the Fukushima incident will be. Discussions are on-going with regard to continued operation and life extension of the existing fleet, new build, and the wider policy issues including technologies needed to address spent fuel storage and ensure energy security, and the related desires to provide sustainable energy systems while at the same time limiting greenhouse gas emissions. The science base for advanced diagnostics and prognostics to support its use in nuclear power plants (NPPs) for active components (pumps, valves etc) has been demonstrated. A challenge is enabling adaption of these technologies for NPP deployment and the validation of the data from these technologies. Advanced diagnostics, monitoring and prognostics applied to passive structures, which in the USA context of longer term operation is up to 80 years, are being researched. Early laboratory work is demonstrating the potential for these methods, although technical challenges remain. It can be expected that there will be an increased need for and use of on-line monitoring for a wide range of both active and passive systems in all types of nuclear power plants.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

MOCMC: Method of Characteristics Moment Closure, a Numerical Method for Covariant Radiation Magnetohydrodynamics

In this work, we present a conservative numerical method for radiation magnetohydrodynamics with frequency-dependent full transport in stationary spacetimes. This method is stable and accurate for both large and small optical depths and radiation pressures. The radiation stress–energy tensor is evolved in flux-conservative form, and closed with a swarm of samples that each transport a multigroup representation of the invariant specific intensity along a null geodesic. In each zone, the enclosed samples are used to efficiently construct a Delaunay triangulation of the unit sphere in the comoving frame, which in turn is used to calculate the Eddington tensor, average source terms, and adaptively refine the sample swarm. Furthermore, radiation four-forces are evaluated in the moment sector in a semi-implicit fashion. The radiative transfer equation is solved in invariant form deterministically for each sample. Since each sample carries a discrete representation of the full spectrum, the cost of evaluating the transport operator is independent of the number of frequency groups, representing a significant reduction of algorithmic complexity for transport in frequency-dependent problems. The major approximation we make in this work is performing scattering in an angle-averaged way. Local adaptivity in samples also makes this scheme more amenable to nonuniform meshes than a traditional Monte Carlo method. We describe the method and present results on a suite of test problems. We find that Method of Characteristics Moment Closure converges at least as ~N -1 , rather than the canonical Monte Carlo N -1/2 , where N is the number of samples per zone.

79 ASTRONOMY AND ASTROPHYSICS↗

Deep Reinforcement Learning for Distribution System Operations: A Tutorial and Survey

Here, the rapid evolution of modern electric power distribution systems into complex networks of interconnected active devices, distributed generation (DG), and storage poses increasing difficulties for system operators. The large-scale integration of distributed energy resources (DERs) and the rapid exchange of measurement data via communication networks present major opportunities for advancing grid operations but also introduce greater uncertainty, higher data dimensionality, more complex network and device models, and challenging control and optimization problems. Deep reinforcement learning (DRL) algorithms are promising in addressing these challenges. However, they have not been effectively adapted for power systems applications, requiring extensive customization for implementation and evaluation. This has resulted in reproducibility challenges and a steep learning curve for researchers new to applying DRL algorithms to the power systems domain. To bridge these gaps, this tutorial aims to serve as a valuable resource for researchers interested in exploring learning-based algorithms to operate active power distribution networks. Specifically, this work presents a generalized process for translating sequential decision-making problems in power distribution systems into Markov decision process (MDP) formulations, illustrated through concrete grid service examples. Additionally, we introduce a simple environment design strategy to develop and evaluate example DRL algorithms for distribution system applications, complete with an included code repository to guide users through environment construction.

24 POWER TRANSMISSION AND DISTRIBUTION↗

SmartTransit.AI: A Dynamic Paratransit and Microtransit Application

New rideshare and shared mobility services have transformed urban mobility in recent years. Such services have the potential to improve efficiency and reduce costs by allowing users to share rides in high-capacity vehicles and vans. Most transit agencies already operate various ridepooling services, including microtransit and paratransit. However, the objectives and constraints for implementing these services vary greatly between agencies and can be challenging. First, off-the-shelf ridepooling formulations must be adapted for real-world conditions and constraints. Second, the lack of modular and reusable software makes it hard to implement and evaluate new ridepooling algorithms and approaches in real-world settings. We demonstrate a modular on-demand public transportation scheduling software for microtransit and paratransit services. The software is aimed at transit agencies looking to incorporate state-of-the-art rideshare and ridepooling algorithms in their everyday operations. We provide management software for dispatchers and mobile applications for drivers and users and conclude with results from the demonstration in Chattanooga, TN.

Pavia, Sophie↗

Influence of Groundwater Extraction Costs and Resource Depletion Limits on Simulated Global Nonrenewable Water Withdrawals Over the Twenty‐First Century

Abstract Future rates of global groundwater depletion will depend on the economic and environmental viability of extracting water from increasingly stressed aquifers. Here we analyze global groundwater depletion by considering these factors explicitly. Global gridded groundwater availability and extraction cost data are aggregated to produce nonrenewable resource supply curves for 235 major river basins and geopolitical regions. These resources are then exposed to dynamically generated demands for water in a fully coupled, multisectoral, global simulation. As groundwater head levels drop, imposing greater capital and operating costs to bring water to the surface, modeled water use sectors are able to deploy a range of supply‐ and demand‐driven adaptive responses. Results demonstrate large sensitivity in global groundwater depletion rates to adjustments in resource exploitability. Extraction costs moderate demands for nonrenewable water substantially, resulting in the onset of a decline in global groundwater depletion rates within the twenty‐first century. New groundwater depletion hot spots may emerge as crop producers abandon overexploited basins and expand croplands in regions with cheaper, more plentiful water resources.

54 ENVIRONMENTAL SCIENCES↗

Fourier-based three-dimensional multistage transformer for aberration correction in multicellular specimens

High-resolution tissue imaging is often compromised by sample-induced optical aberrations that degrade resolution and contrast. Although wavefront sensor-based adaptive optics (AO) can measure these aberrations, such hardware solutions are typically complex, expensive to implement and slow when serially mapping spatially varying aberrations across large fields of view. Here we introduce AOViFT (adaptive optical vision Fourier transformer)—a machine learning-based aberration sensing framework built around a three-dimensional multistage vision transformer that operates on Fourier domain embeddings. AOViFT infers aberrations and restores diffraction-limited performance in puncta-labeled specimens with substantially reduced computational cost, training time and memory footprint compared to conventional architectures or real-space networks. We validated AOViFT on live gene-edited zebrafish embryos, demonstrating its ability to correct spatially varying aberrations using either a deformable mirror or postacquisition deconvolution. By eliminating the need for the guide star and wavefront sensing hardware and simplifying the experimental workflow, AOViFT lowers technical barriers for high-resolution volumetric microscopy across diverse biological samples.

Alshaabi, Thayer [Howard Hughes Medical Institute,↗

A framework for data-driven solution and parameter estimation of PDEs using conditional generative adversarial networks

We employ and adapt the image-to-image translation concept based on conditional generative adversarial networks (cGAN) for learning a forward and an inverse solution operator of partial differential equations (PDEs). We focus on steady-state solutions of coupled hydromechanical processes in heterogeneous porous media and present the parameterization of the spatially heterogeneous coefficients, which is exceedingly difficult using standard reduced-order modeling techniques. We show that our framework provides a speed-up of at least 2,000 times compared to a finite-element solver and achieves a relative root-mean-square error (r.m.s.e.) of less than 2% for forward modeling. For inverse modeling, the framework estimates the heterogeneous coefficients, given an input of pressure and/or displacement fields, with a relative r.m.s.e. of less than 7%, even for cases where the input data are incomplete and contaminated by noise. The framework also provides a speed-up of 120,000 times compared to a Gaussian prior-based inverse modeling approach while also delivering more accurate results.

97 MATHEMATICS AND COMPUTING↗

COSMOLOGICAL ADAPTIVE MESH REFINEMENT MAGNETOHYDRODYNAMICS WITH ENZO

In this work, we present EnzoMHD, the extension of the cosmological code Enzo to include the effects of magnetic fields through the ideal magnetohydrodynamics approximation. We use a higher order Godunov method for the computation of interface fluxes. We use two constrained transport methods to compute the electric field from those interface fluxes, which simultaneously advances the induction equation and maintains the divergence of the magnetic field. A second-order divergence-free reconstruction technique is used to interpolate the magnetic fields in the block-structured adaptive mesh refinement framework already extant in Enzo. This reconstruction also preserves the divergence of the magnetic field to machine precision. We use operator splitting to include gravity and cosmological expansion. We then present a series of cosmological and non-cosmological test problems to demonstrate the quality of solution resulting from this combination of solvers.

79 ASTRONOMY AND ASTROPHYSICS↗

Renormalization of heavy-light currents in moving nonrelativistic QCD

Heavy-light decays such as B{yields}{pi}l{nu}, B{yields}K*{gamma} and B{yields}K{sup (*)}ll can be used to constrain the parameters of the standard model and in indirect searches for new physics. While the precision of experimental results has improved over the last years this has still to be matched by equally precise theoretical predictions. The calculation of heavy-light form factors is currently carried out in lattice QCD. Because of its small Compton wavelength we discretize the heavy quark in an effective nonrelativistic theory. By formulating the theory in a moving frame of reference discretization errors in the final state are reduced at large recoil. Over the last years the formalism has been improved and tested extensively. Systematic uncertainties are reduced by renormalizing the m(oving) NRQCD action and heavy-light decay operators. The theory differs from QCD only for large loop momenta at the order of the lattice cutoff and the calculation can be carried out in perturbation theory as an expansion in the strong coupling constant. In this paper we calculate the one-loop corrections to the heavy-light vector and tensor operator. Because of the complexity of the action the generation of lattice Feynman rules is automated and loop integrals are solved by the adaptive Monte Carlo integrator VEGAS. We discuss the infrared and ultraviolet divergences in the loop integrals both in the continuum and on the lattice. The light quarks are discretized in the ASQTad and highly improved staggered quark (HISQ) action; the formalism is easily extended to other quark actions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Enhancing the Performance of Assisted Execution Runtime Systems through Hardware/Software Techniques

To meet the expected performance, future exascale systems will require programmers to increase the level of parallelism of their applications. Novel programming models simplify parallel programming at the cost of increasing runtime overheard. Assisted execution models have the potential of reducing this overhead but they generally also reduce processor utilization. We propose an integrated hardware/software solution that automatically partition hardware resources between application and auxiliary threads. Each system level performs well-defined tasks efficiently: 1) the runtime system is enriched with a mechanism that automatically detects computing power requirements of running threads and drives the hardware actuators; 2) the hardware enforces dynamic resource partitioning; 3) the operating system provides an efficient interface between the runtime system and the hardware resource allocation mechanism. As a test case, we apply this adaptive approach to STM2, an software transactional memory system that implements the assisted execution model. We evaluate the proposed adaptive solution on an IBMPOWER7 system using Eigenbench and STAMP benchmark suite. Results show that our approach performs equal or better than the original STM2 and achieves up to 65% and 86% performance improvement for Eigenbench and STAMP applications, respectively.

Kestor, Gokcen↗

Diventi

Diventi provides a write optimized index of network logs such as netflow or zeek. The program will ingest various network traffic logs (zeek connection logs or NetflowV5, v9 or IPFix logs) and index each IP address and time of its activity. This database can then be queried for timely tracking of network activity. It currently uses the ft-index library from Percona but is adaptable to any key-value store that supports the Berkeley DB api.SAND2019-15037 M Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Kroeger, Thomas↗

BASS

SAND2026-17001O BASS implements Bayesian Adaptive Spline Surfaces in MATLAB and serves as a surrogate model for regression applications. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-NA0003525.

Tucker, J. Derek [Sandia National Lab. (SNL-CA), L↗

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↗