Scalable uncertainty quantification for deep operator networks using randomized priors
Not provided.
Engineering↗
SEARCH · Search NASA
Results for “parallel, 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.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Non-local large-strain FFT-based formulation and its application to interface-dominated plasticity of nano-metallic laminates
Here, this paper presents a novel formulation and its robust numerical implementation of strain-gradient (SG) crystal plasticity within a large-strain (LS) elasto-viscoplastic (EVP) fast Fourier transform (FFT)-based micromechanical model. The resulting non-local SG-LS-EVPFFT formulation is used to model and understand the process of kink band formation during layer-parallel compression of nano-metallic laminates (NMLs). NMLs are layered composites with nanoscale thicknesses, thus requiring consideration of the interaction between dislocations and interfaces within the micromechanical model. The length-scale parameter of the SG model is calibrated by simulating a double pile-up and comparing predictions to analytical solution. This required new expressions for the defect energy, resulting in more accurate double pile-up predictions. The calibrated SG-LS-EVPFFT model is then used to simulate layer-parallel compression of copper-niobium NML. Formation of kink bands is predicted, and the model is used to rationalize the microscopic mechanisms enabling the formation process. It is found that accumulation of dislocations at interfaces leads to activation of layer-parallel slip, which in turn leads to kink band formation.
36 MATERIALS SCIENCE↗
Parallel quantum computing simulations via quantum accelerator platform virtualization
Quantum circuit execution is a central task in quantum computation. Due to inherent quantum-mechanical constraints, quantum computing workflows often involve a considerable number of independent measurements over a large set of slightly different quantum circuits. Here we discuss a simple model for parallelizing such quantum circuit executions that is based on introducing a large array of virtual quantum processing units (mapped to HPC nodes in our case) as a parallel quantum computing platform. Implemented within the XACC framework, the model can readily take advantage of its backend-agnostic features, enabling parallel quantum computing/simulation over any target backend supported by XACC. We illustrate the performance of this approach by demonstrating strong scaling in two pertinent domain science problems, namely in computing the gradients for the multi-contracted variational quantum eigensolver and in data-driven quantum circuit learning, where we vary the number of qubits and the number of circuit layers. Here, the latter simulation leverages the cuQuantum library to run efficiently on GPU-accelerated HPC platforms.
97 MATHEMATICS AND COMPUTING↗
Stochastic evaluation of four-component relativistic second-order many-body perturbation energies: A potentially quadratic-scaling correlation method
A second-order many-body perturbation correction to the relativistic Dirac-Hartree-Fock energy is evaluated stochastically by integrating 13-dimensional products of four-component spinors and Coulomb potentials. The integration in the real space of electron coordinates is carried out by the Monte Carlo (MC) method with the Metropolis sampling, whereas the MC integration in the imaginary-time domain is performed by the inverse-CDF (cumulative distribution function) method. The computational cost to reach a given relative statistical error for spatially compact but heavy molecules is observed to be no worse than cubic and possibly quadratic with the number of electrons or basis functions. This is a vast improvement over the quintic scaling of the conventional, deterministic second-order many-body perturbation method. The algorithm is also easily and efficiently parallelized with demonstrated 92% strong scalability going from 64 to 4096 processors for a fixed job size.
74 ATOMIC AND MOLECULAR PHYSICS↗
A High Performance Block Eigensolver for Nuclear Configuration Interaction Calculations
As on-node parallelism increases and the performance gap between the processor and the memory system widens, achieving high performance in large-scale scientific applications requires an architecture-aware design of algorithms and solvers. We focus on the eigenvalue problem arising in nuclear Configuration Interaction (CI) calculations, where a few extreme eigenpairs of a sparse symmetric matrix are needed. Here, we consider a block iterative eigensolver whose main computational kernels are the multiplication of a sparse matrix with multiple vectors (SpMM), and tall-skinny matrix operations. We then present techniques to significantly improve the SpMM and the transpose operation SpMM T by using the compressed sparse blocks (CSB) format. We achieve 3-4× speedup on the requisite operations over good implementations with the commonly used compressed sparse row (CSR) format. We develop a performance model that allows us to correctly estimate the performance of our SpMM kernel implementations, and we identify cache bandwidth as a potential performance bottleneck beyond DRAM. We also analyze and optimize the performance of LOBPCG kernels (inner product and linear combinations on multiple vectors) and show up to 15× speedup over using high performance BLAS libraries for these operations. The resulting high performance LOBPCG solver achieves 1.4× to 1.8× speedup over the existing Lanczos solver on a series of CI computations on high-end multicore architectures (Intel Xeons). We also analyze the performance of our techniques on an Intel Xeon Phi Knights Corner (KNC) processor.
97 MATHEMATICS AND COMPUTING↗
Mobiliti v1.0
Mobiliti is a software platform designed to emulate the dynamics of a regional transportation road network. It is built on open-source software that provides parallel discrete-event simulation. The software is transformative in the area of transportation network simulation because of the geospatial scale and fidelity of the network model and the computational time it takes to model a full day of travel demand. For example, it runs a simulation of the entire San Francisco Bay Area, with a network model of ~1M links and a population that completes ~19M trips in ~5 minutes. This scale of simulation has not been attempted with existing simulation models due to the complexity of the model and the computational time it would take to complete. The intent of the software is to create a digital twin capability for cities to evaluate consequences of infrastructure or policy changes on road network dynamics.
Macfarlane, Jane↗
A mechanical reduced order model for elastomeric 3D printed architectures
Direct ink writing of silicone elastomers enables printing with precise control of porosity and mechanical properties of ordered cellular solids, suitable for shock absorption and stress mitigation applications. With the ability to manipulate structure and feedstock stiffness, the design space becomes challenging to parse to obtain a solution producing a desired mechanical response. We derive an analytical design approach for a specific architecture. Results from finite element simulations and quasi-static mechanical tests of two different parallel strand architectures were analyzed to understand the structure-property relationships under uniaxial compression. Combining effective stiffness-density scaling with least squares optimization of the stress responses yielded general response curves parameterized by resin modulus and strand spacing. An analytical expression of these curves serves as a reduced order model, which, when optimized, provides a rapid design capability for filament-based 3D printed structures. Finally, as a demonstration, the optimal design of a face-centered tetragonal architecture is computed that satisfies prescribed minimum and maximum load constraints.
36 MATERIALS SCIENCE↗
Universal Design Curves for Elastomeric Direct Ink-Writing Architectures
Direct ink writing of silicone elastomers enables printing with precise control of porosity and mechan- ical properties of ordered cellular solids, suitable for shock absorption and stress mitigation applications. With the ability to manipulate structure and feedstock stiffness, the design space becomes challenging to parse to obtain a solution producing a desired mechanical response. Here we derive an analytical design approach for a specific architecture. Results from finite element simulations and quasi-static mechanical tests of two different parallel strand architectures were analyzed to understand the structure-property relationships under compressive strain. Combining effective stiffness-density scaling with least squares optimization of the stress responses yielded general response curves parameterized by resin modulus and strand spacing. An analytical expression of these curves serves as a reduced order model, which, when optimized, provides a rapid design capability for filament-based 3D printed structures. As a demonstration, the optimal design of a face-centered tetragonal architecture is computed that satisfies prescribed minimum and maximum load constraints.
36 MATERIALS SCIENCE↗
Transient and Steady-State Inverse Problems in Sierra/Aria
Inverse problems arise in a wide range of applications, whenever unknown model parameters cannot be measured directly. Instead, the unknown parameters are estimated using experimental data and forward simulations. Thermal inverse problems, such as material characterization problems, are often large-scale and transient. Therefore, they require intrusive adjoint-based gradient implementations in order to be solved efficiently. The capability to solve large-scale transient thermal inverse problems using an adjoint-based approach was recently implemented in SNL Sierra Mechanics, a massively parallel capable multiphysics code suite. This report outlines the theory, optimization formulation, and path taken to implement thermal inverse capabilities in Sierra within a unit test framework. The capability utilizes Sierra/Aria and Sierra/Fuego data structures, the Rapid Optimization Library, and an interface to the Sierra/InverseOpt library. The existing Sierra/Aria time integrator is leveraged to implement a time-dependent adjoint solver.
97 MATHEMATICS AND COMPUTING↗
Electron Drift Kinetics in Space and in the Lab [Slides]
In a magnetized plasma, dynamics are fundamentally split between the directions parallel and perpendicular to the magnetic field. Generally, gyration about the magnetic field will occur on smaller spatial and time scales than the bulk dynamics of the system. From Hamiltonian mechanics, a periodic motion on a time scale much shorter than the other system dynamics leads to an adiabatic invariant. The adiabatic invariant for this motion is the magnetic moment $\mu=\frac{mv⟂^2}{2B}$, which will be approximately conserved.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Quantum Theory, Quantum Materials, Quantum Computing
The Sanibel Symposium series is renowned amongst materials theorists, quantum chemists, and condensed matter physicists as meetings driving progress on theory, mod eling, and simulation of materials and their molecular and nano-scale constituents. The Symposia are highly unusual (perhaps unique) in their priority emphasis on theory and computation, in having no parallel sessions, in cultivating well-attended Hot-Topic contributed oral sessions, and accessible poster sessions. These provide highly visible, influential platforms for cross-fertilization among specialist investigators, hence are strong contributors to the advance of quantum information sciences (QIS) research of strategic importance to the Office of Basic Energy Sciences (BES). As part of a five-year plan to highlight QIS challenges and opportunities and foster progress on them, each of the pre ceding three Sanibel Symposia had a thematic focus, Quantum Theory, Quantum Materi als, Quantum Computing, as a major program component. Emphasis was on quantum materials and their molecular constituents. The award for 2024 was for year four of that sustained thematic focus.
36 MATERIALS SCIENCE↗
A GENERALIZED TWO-COMPONENT MODEL OF SOLAR WIND TURBULENCE AND AB INITIO DIFFUSION MEAN-FREE PATHS AND DRIFT LENGTHSCALES OF COSMIC RAYS
We extend a two-component model for the evolution of fluctuations in the solar wind plasma so that it is fully three-dimensional (3D) and also coupled self-consistently to the large-scale magnetohydrodynamic equations describing the background solar wind. The two classes of fluctuations considered are a high-frequency parallel-propagating wave-like piece and a low-frequency quasi-two-dimensional component. For both components, the nonlinear dynamics is dominanted by quasi-perpendicular spectral cascades of energy. Driving of the fluctuations by, for example, velocity shear and pickup ions is included. Numerical solutions to the new model are obtained using the Cronos framework, and validated against previous simpler models. Comparing results from the new model with spacecraft measurements, we find improved agreement relative to earlier models that employ prescribed background solar wind fields. Finally, the new results for the wave-like and quasi-two-dimensional fluctuations are used to calculate ab initio diffusion mean-free paths and drift lengthscales for the transport of cosmic rays in the turbulent solar wind.
79 ASTRONOMY AND ASTROPHYSICS↗
Welded solar cell interconnection
The efficiency of the welding of solar-cell interconnects is compared with the efficiency of soldering such interconnects, and the cases in which welding may be superior are examined. Emphasis is placed on ultrasonic welding; attention is given to the solar-cell welding machine, the application of the welding process to different solar-cell configurations, producibility, and long-life performance of welded interconnects. Much of the present work has been directed toward providing increased confidence in the reliability of welding using conditions approximating those that would occur with large-scale array production. It is concluded that there is as yet insufficient data to determine which of three methods (soldering, parallel gap welding, and ultrasonic welding) provides the longest-duration solar panel life.
Stofel, E. J.↗
Compositional changes, time and density variations in the magnetosphere associated with Birkeland currents and particle acceleration
Birkeland currents, parallel electric fields and plasma instabilities often occur together in time and space and play an important role in large scale plasma motions between the ionosphere and magnetosphere and along the magnetic field in the magnetosphere. The results of the plasma movements are large density and composition variations in the ionosphere and magnetosphere. Observations from ISIS-2 at 1400 km altitude show large densities with heavy ions dominating in regions with upward Birkeland currents, and low densities and light ions in regions with downward currents. Observations from ISEE-1 in field aligned current regions at 10,000 to 15,000 km altitude show transverse heating of protons and oxygen ions to 250 eV. Because of the different mobility of the protons and oxygen ions the proton flow is important in the beginning of the events but later the outflow becomes almost pure oxygen. Similarly ISEE-1 observations of outgoing field ion beams at 10,000 to 15,000 km altitude show time variations in the H+/O+ ratios and a dominance of O+ later in the events.
Ungstrup, E.↗
Numerical simulation of vortex unsteadiness on slender bodies of revolution at large incidence
Time-accurate, fine-grid Navier-Stokes solutions were obtained for flow over a slender ogive-cylinder body of revolution at angles of attack ranging from 10 deg to 40 deg. The results indicate the progressive growth of crossflow separation and the development of the leeward side vortex pattern with increasing incidence. The computed flows show good agreement with experimental measurements. As the angle of attack was increased, the flows become less damped, and at 40 deg a nonsteady flow exhibiting self-sustained fluctuations was observed. The nonsteadiness was linked to the presence of small-scale three-dimensional vortices moving along the primary surfaces of crossflow separation. The behavior of the fluctuations with incidence parallels the trend observed in experiments.
Schiff, Lewis B.↗
Probabilistic structural mechanics research for parallel processing computers
Aerospace structures and spacecraft are a complex assemblage of structural components that are subjected to a variety of complex, cyclic, and transient loading conditions. Significant modeling uncertainties are present in these structures, in addition to the inherent randomness of material properties and loads. To properly account for these uncertainties in evaluating and assessing the reliability of these components and structures, probabilistic structural mechanics (PSM) procedures must be used. Much research has focused on basic theory development and the development of approximate analytic solution methods in random vibrations and structural reliability. Practical application of PSM methods was hampered by their computationally intense nature. Solution of PSM problems requires repeated analyses of structures that are often large, and exhibit nonlinear and/or dynamic response behavior. These methods are all inherently parallel and ideally suited to implementation on parallel processing computers. New hardware architectures and innovative control software and solution methodologies are needed to make solution of large scale PSM problems practical.
Sues, Robert H.↗
Particle acceleration in flares
Particle acceleration is intrinsic to the primary energy release in the impulsive phase of solar flares, and we cannot understand flares without understanding acceleration. New observations in soft and hard X-rays, gamma-rays and coherent radio emissions are presented, suggesting flare fragmentation in time and space. X-ray and radio measurements exhibit at least five different time scales in flares. In addition, some new observations of delayed acceleration signatures are also presented. The theory of acceleration by parallel electric fields is used to model the spectral shape and evolution of hard X-rays. The possibility of the appearance of double layers is further investigated.
Benz, Arnold O.↗
Parallel-vector out-of-core equation solver for computational mechanics
A parallel/vector out-of-core equation solver is developed for shared-memory computers, such as the Cray Y-MP machine. The input/ output (I/O) time is reduced by using the a synchronous BUFFER IN and BUFFER OUT, which can be executed simultaneously with the CPU instructions. The parallel and vector capability provided by the supercomputers is also exploited to enhance the performance. Numerical applications in large-scale structural analysis are given to demonstrate the efficiency of the present out-of-core solver.
Qin, J.↗