Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “compute intensive problems”

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 289 records · Page 16

GPU-enabled extreme-scale turbulence simulations: Fourier pseudo-spectral algorithms at the exascale using OpenMP offloading

Fourier pseudo-spectral methods for nonlinear partial differential equations are of wide interest in many areas of advanced computational science, including direct numerical simulation of three-dimensional (3-D) turbulence governed by the Navier-Stokes equations in fluid dynamics. This paper presents a new capability for simulating turbulence at a new record resolution up to 35 trillion grid points, on the world's first exascale computer, Frontier, comprising AMD MI250x GPUs with HPE's Slingshot interconnect and operated by the US Department of Energy's Oak Ridge Leadership Computing Facility (OLCF). Key programming strategies designed to take maximum advantage of the machine architecture involve performing almost all computations on the GPU which has the same memory capacity as the CPU, performing all-to-all communication among sets of parallel processes directly on the GPU, and targeting GPUs efficiently using OpenMP offloading for intensive number-crunching including 1-D Fast Fourier Transforms (FFT) performed using AMD ROCm library calls. With 99% of computing power on Frontier being on the GPU, leaving the CPU idle leads to a net performance gain via avoiding the overhead of data movement between host and device except when needed for some I/O purposes. Memory footprint including the size of communication buffers for MPI_ALLTOALL is managed carefully to maximize the largest problem size possible for a given node count. Detailed performance data including separate contributions from different categories of operations to the elapsed wall time per step are reported for five grid resolutions, from 2048 3 on a single node to 32768 3 on 4096 or 8192 nodes out of 9408 on the system. Both 1D and 2D domain decompositions which divide a 3D periodic domain into slabs and pencils respectively are implemented. The present code suite (labeled by the acronym GESTS, GPUs for Extreme Scale Turbulence Simulations) achieves a figure of merit (in grid points per second) exceeding goals set in the Center for Accelerated Application Readiness (CAAR) program for Frontier. The performance attained is highly favorable in both weak scaling and strong scaling, with notable departures only for 2048 3 where communication is entirely intra-node, and for 32768 3 , where a challenge due to small message sizes does arise. Communication performance is addressed further using a lightweight test code that performs all-to-all communication in a manner matching the full turbulence simulation code. Performance at large problem sizes is affected by both small message size due to high node counts as well as dragonfly network topology features on the machine, but is consistent with official expectations of sustained performance on Frontier. Overall, although not perfect, the scalability achieved at the extreme problem size of 32768 3 (and up to 8192 nodes — which corresponds to hardware rated at just under 1 exaflop/sec of theoretical peak computational performance) is arguably better than the scalability observed using prior state-of-the-art algorithms on Frontier's predecessor machine (Summit) at OLCF. New science results for the study of intermittency in turbulence enabled by this code and its extensions are to be reported separately in the near future.

3D fast Fourier transform↗

CSRI Summer Proceedings 2020

The Computer Science Research Institute (CSRI) brings university faculty and students to Sandia for focused collaborative research on Department of Energy (DOE) computer and computational science problems. The institute provides an opportunity for university researchers to learn about problems in computer and computational science at DOE laboratories. Participants conduct leading-edge research, interact with scientists and engineers at the laboratories, and help transfer results of their research to programs at the labs. Some specific CSRI research interest areas are: scalable solvers, optimization, adaptivity and mesh refinement, graph-based, discrete, and combinatorial algorithms, uncertainty estimation, mesh generation, dynamic load-balancing, virus and other malicious-code defense, visualization, scalable cluster computers, data-intensive computing, environments for scalable computing, parallel input/output, advanced architectures, and theoretical computer science. The CSRI Summer Program is organized by CSRI and typically includes the organization of a weekly seminar series and the publication of a summer proceedings. In 2020, the CSRI summer program was executed completely virtually; all student interns worked from home, due to the COVID-19 pandemic.

97 MATHEMATICS AND COMPUTING↗

Acoustic Wave Propagation Modeling by a Two-dimensional Finite-difference Summation-by-parts Algorithm

Acoustic waveform modeling is a computationally intensive task and full three-dimensional simulations are often impractical for some geophysical applications such as long-range wave propagation and high-frequency sound simulation. In this study, we develop a two-dimensional high-order accurate finite-difference code for acoustic wave modeling. We solve the linearized Euler equations by discretizing them with the sixth order accurate finite difference stencils away from the boundary and the third order summation-by-parts (SBP) closure near the boundary. Non-planar topographic boundary is resolved by formulating the governing equation in curvilinear coordinates following the interface. We verify the implementation of the algorithm by numerical examples and demonstrate the capability of the proposed method for practical acoustic wave propagation problems in the atmosphere.

58 GEOSCIENCES↗

A comparison of Gaussian processes and neural networks for computer model emulation and calibration

The Department of Energy relies on complex physics simulations for prediction in domains like cosmology, nuclear theory, and materials science. These simulations are often extremely computationally intensive, with some requiring days or weeks for a single simulation. In order to assure their accuracy, these models are calibrated against observational data in order to estimate inputs and systematic biases. Because of their great computational complexity, this process typically requires the construction of an emulator, a fast approximation to the simulation. In this paper, two emulator approaches are compared: Gaussian process regression and neural networks. Their emulation accuracy and calibration performance on three real problems of Department of Energy interest is considered. On these problems, the Gaussian process emulator tends to be more accurate with narrower, but still well-calibrated uncertainty estimates. The neural network emulator is accurate, but tends to have large uncertainty on its predictions. Finally, as a result, calibration with the Gaussian process emulator produces more constrained posteriors that still perform well in prediction.

97 MATHEMATICS AND COMPUTING↗

QA Issues for Computer-Controlled Treatment Delivery: This Is Not Your Old R/V System Any More{exclamation_point}

State-of-the-art radiotherapy treatment delivery has changed dramatically during the past decade, moving from manual individual field setup and treatment to automated computer-controlled delivery of complex treatments, including intensity-modulated radiotherapy and other similarly complex delivery strategies. However, the quality assurance methods typically used to ensure treatment is performed precisely and correctly have not evolved in a similarly dramatic way. This paper reviews the old manual treatment process and use of record-and-verify systems, and describes differences with modern computer-controlled treatment delivery. The process and technology used for computer-controlled treatment delivery are analyzed in terms of potential (and actual) problems, as well as relevant published guidance on quality assurance. The potential for improved quality assurance for computer-controlled delivery is discussed.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Polarization-dependent atomic dipole traps behind a circular aperture for neutral-atom quantum computing

The neutral-atom quantum computing community has successfully implemented almost all necessary steps for constructing a neutral-atom quantum computer. We present computational results of a study aimed at solving the remaining problem of creating a quantum memory with individually addressable sites for quantum computing. The basis of this quantum memory is the diffraction pattern formed by laser light incident on a circular aperture. Very close to the aperture, the diffraction pattern has localized bright and dark spots that can serve as red-detuned or blue-detuned atomic dipole traps. These traps are suitable for quantum computing even for moderate laser powers. In particular, for moderate laser intensities ({approx}100 W/cm{sup 2}) and comparatively small detunings ({approx}1000-10 000 linewidths), trap depths of {approx}1 mK and trap frequencies of several to tens of kilohertz are achieved. Our results indicate that these dipole traps can be moved by tilting the incident laser beams without significantly changing the trap properties. We also explored the polarization dependence of these dipole traps. We developed a code that calculates the trapping potential energy for any magnetic substate of any hyperfine ground state of any alkali-metal atom for any laser detuning much smaller than the fine-structure splitting for any given electric field distribution. We describe details of our calculations and include a summary of different notations and conventions for the reduced matrix element and how to convert it to SI units. We applied this code to these traps and found a method for bringing two traps together and apart controllably without expelling the atoms from the trap and without significant tunneling probability between the traps. This approach can be scaled up to a two-dimensional array of many pinholes, forming a quantum memory with single-site addressability, in which pairs of atoms can be brought together and apart for two-qubit gates for quantum computing.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Z2Pack: Numerical implementation of hybrid Wannier centers for identifying topological materials

The intense theoretical and experimental interest in topological insulators and semimetals has established band structure topology as a fundamental material property. Consequently, identifying band topologies has become an important, but often challenging, problem, with no exhaustive solution at the present time. Here, we compile a series of techniques, some previously known, that allow for a solution to this problem for a large set of the possible band topologies. The method is based on tracking hybrid Wannier charge centers computed for relevant Bloch states, and it works at all levels of materials modeling: continuous k ∙ p models, tight-binding models, and ab initio calculations. We apply the method to compute and identify Chern, $\mathbb{Z}$ 2 , and crystalline topological insulators, as well as topological semimetal phases, using real material examples. Moreover, we provide a numerical implementation of this technique (the Z2Pack software package) that is ideally suited for high-throughput screening of materials databases for compounds with nontrivial topologies. We expect that our work will allow researchers to (a) identify topological materials optimal for experimental probes, (b) classify existing compounds, and (c) reveal materials that host novel, not yet described, topological states.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

SharP

The pre-exascale systems are expected to have a significant amount of hierarchical and heterogeneous on-node memory, and this trend of system architecture in extreme-scale systems is expected to continue into the exascale era. along with hierarchical-heterogeneous memory, the system typically has a high-performing network ad a compute accelerator. This system architecture is not only effective for running traditional High Performance Computing (HPC) applications (Big-Compute), but also for running data-intensive HPC applications and Big-Data applications. As a consequence, there is a growing desire to have a single system serve the needs of both Big-Compute and Big-Data applications. Though the system architecture supports the convergence of the Big-Compute and Big-Data, the programming models and software layer have yet to evolve to support either hierarchical-heterogeneous memory systems or the convergence. A programming abstraction to address this problem. The programming abstraction is implemented as a software library and runs on pre-exascale and exascale systems supporting current and emerging system architecture. Using distributed data-structures as a central concept, it provides (1) a simple, usable, and portable abstraction for hierarchical-heterogeneous memory and (2) a unified programming abstraction for Big-Compute and Big-Data applications.

Venkata, Manjunath Gorentla↗

GPU accelerated fully space and time resolved numerical simulations of self-focusing laser beams in SBS-active media

A computer cluster equipped with Graphics Processing Units (GPUs) is used for simulating nonlinear optical wave packets undergoing Kerr self-focusing and stimulated Brillouin scattering in fused silica. We first recall the model equations in full (3+1) dimensions. These consist of two coupled nonlinear Schrödinger equations for counterpropagating optical beams closed with a source equation for light-induced acoustic waves seeded by thermal noise. Compared with simulations on a conventional cluster of Central Processing Units (CPUs), GPU-based computations allow us to use a significant (16 times) larger number of mesh points within similar computation times. Reciprocally, simulations employing the same number of mesh points are between 3 and 20 times faster on GPUs than on the same number of classical CPUs. Performance speedups close to 45 are reported for isolated functions evaluating, e.g., the optical nonlinearities. Since the field intensities may reach the ionization threshold of silica, the action of a defocusing electron plasma is also addressed.

97 MATHEMATICS AND COMPUTING↗

Community-driven development for computational biology at Sprints, Hackathons and Codefests

Background: Computational biology comprises a wide range of technologies and approaches. Multiple technologies can be combined to create more powerful workflows if the individuals contributing the data or providing tools for its interpretation can find mutual understanding and consensus. Much conversation and joint investigation are required in order to identify and implement the best approaches. Traditionally, scientific conferences feature talks presenting novel technologies or insights, followed up by informal discussions during coffee breaks. In multi-institution collaborations, in order to reach agreement on implementation details or to transfer deeper insights in a technology and practical skills, a representative of one group typically visits the other. However, this does not scale well when the number of technologies or research groups is large. Conferences have responded to this issue by introducing Birds-of-a-Feather (BoF) sessions, which offer an opportunity for individuals with common interests to intensify their interaction. However, parallel BoF sessions often make it hard for participants to join multiple BoFs and find common ground between the different technologies, and BoFs are generally too short to allow time for participants to program together. Results: This report summarises our experience with computational biology Codefests, Hackathons and Sprints, which are interactive developer meetings. They are structured to reduce the limitations of traditional scientific meetings described above by strengthening the interaction among peers and letting the participants determine the schedule and topics. These meetings are commonly run as loosely scheduled “unconferences” (self-organized identification of participants and topics for meetings) over at least two days, with early introductory talks to welcome and organize contributors, followed by intensive collaborative coding sessions. We summarise some prominent achievements of those meetings and describe differences in how these are organised, how their audience is addressed, and their outreach to their respective communities. Conclusions: Hackathons, Codefests and Sprints share a stimulating atmosphere that encourages participants to jointly brainstorm and tackle problems of shared interest in a self-driven proactive environment, as well as providing an opportunity for new participants to get involved in collaborative projects.

59 BASIC BIOLOGICAL SCIENCES↗

A computationally efficient P 1 radiation model for modern combustion systems utilizing pre-conditioned conjugate gradient methods

The iterative convergence of the P 1 radiation model can be slow in optically thin scenarios when employing classical iterative methods. In order to remedy this shortcoming, an in-house P 1 radiation model was interfaced with high performance, scalable, linear solver libraries. Next, the accuracies of P 1 radiation model calculations was assessed by comparing its predictions against discrete ordinates (DO) model calculations for prototypical problems representative of modern combustion systems. Corresponding benchmark results were also included for comparison. Utilizing Pre-Conditioners (PC) to the Conjugate Gradients (CG) method, the convergence time of the P 1 radiation model reduced by a factor of 30 for modest problem sizes and a factor of 70 for larger sized problems when compared against classical Gauss Seidel sweeps. Additionally, PC provided 50% computational savings compared to employing CG in a standalone mode. The P 1 model calculation times were about 25–30% of the DO model calculation time. The time to solution also scaled linearly with an increase in problem size. The weighted sum of gray gases model employed in this study in conjunction with the P 1 model provided good agreement against benchmark data with L 2 error norms (defined relative to corresponding DO calculations) improving when isotropic intensities were prevalent.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Parallel Implementation of Power System Dynamic Simulation

Dynamic simulation of power system transient stability is important for planning, monitoring, operation, and control of electrical power systems. However, modeling the system dynamics and network involves the computationally intensive time-domain solution of numerous differential and algebraic equations (DAE). This results in a transient stability implementation that may not maintain the real-time constraints of an online security assessment. This paper presents a parallel implementation of the dynamic simulation on a high-performance computing (HPC) platform using parallel simulation algorithms and computation architectures. It enables the simulation to run even faster than real time, enabling the “look-ahead” capability of upcoming stability problems in the power grid.

dynamic simulation↗

Summer Proceedings 2019

The Computer Science Research Institute (CSRI) brings university faculty and students to Sandia for focused collaborative research on Department of Energy (DOE) computer and computational science problems. The institute provides an opportunity for university researchers to learn about problems in computer and computational science at DOE laboratories. Participants conduct leading-edge research, interact with scientists and engineers at the laboratories, and help transfer results of their research to programs at the labs. Some specific CSRI research interest areas are: scalable solvers, optimization, adaptivity and mesh refinement, graph-based, discrete, and combinatorial algorithms, uncertainty estimation, mesh generation, dynamic load-balancing, virus and other malicious-code defense, visualization, scalable cluster computers and heterogeneous computers, data-intensive computing, environments for scalable computing, parallel input/output, advanced architectures, and theoretical computer science. The CSRI Summer Program includes the organization of a weekly seminar series and the publication of this summer proceedings.

97 MATHEMATICS AND COMPUTING↗

An extended Vlasov–Fokker–Planck approach for kinetic simulations of laser plasmas

Vlasov–Fokker–Planck simulation codes occupy an important niche in modeling laser-produced plasmas, since they are well suited to studying the effect of collisions on electron kinetic phenomena, especially energy transport. One of the most important elements of energy transport is the absorption of laser light by the plasma; however, simulating this in detail requires resolving oscillations of the laser light, whose characteristic timescale is orders of magnitude shorter than the simulation time needed to study transport physics. For this reason, most Vlasov–Fokker–Planck codes used to study electron transport in laser plasmas rely on simplified models of the laser–plasma coupling. Their underlying assumptions nominally preclude their use for modeling laser light having short-scale structure in space or time, such as broadband lasers. In this work, we derive a more general computational framework suitable for arbitrarily structured laser fields. Furthermore, our approach is based on an extended set of Vlasov–Fokker–Planck equations that separately solve for the low- and high-frequency plasma response. We implement these extended Vlasov–Fokker–Planck equations in the spherical harmonic code K2 and demonstrate the performance of the method on several laser absorption test problems, with particular attention to the judicious selection of time steps, time integrators, and spherical harmonic truncation, according to the intensity and spectrum of the laser light under consideration. Comparison with the widely used Langdon absorption operator shows the Langdon operator performs remarkably well for predicting laser heating in the simple cases considered here, even in situations that would seem to violate its underlying assumptions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Direct Numerical Simulation of the Turbulent Flow Through PWR Spacer Grid and Mixing Vanes

Direct numerical simulation (DNS) has been regarded as a reliable data source for the development and validation of turbulence models along with experiments. The realization of DNS usually involves a very fine mesh that should be able to resolve all relevant turbulence scales down to Kolmogorov scale. As the most computationally expensive approach compared to other CFD techniques, DNS applications used to be limited to flow studies at very low Reynolds numbers. Thanks to the tremendous growth of computing power over the past decades, the simulation capability of DNS has now started overlapping with some of the most challenging engineering problems. One of those examples in nuclear engineering is the turbulent coolant flow inside reactor cores. In Light Water Reactor (LWR) cores, the nuclear fuel rods are arranged together as rod bundles in a triangular or square pattern. The fuel rod bundles are supported by the spacer grids. The mixing vanes are also installed to promote the turbulence intensity in coolant flow. An understanding of spacer grid and mixing vanes effect is important because it has a direct impact on the heat removal efficiency from the fuel rod surface. In the presented work, the DNS of turbulent flows are performed in three different types of geometries: (a) a regular Pressurized Water Reactor (PWR) subchannel; (b) a special PWR subchannel with reduced size spacer grid and mixing vanes (SGMV); (c) a special PWR subchannel with reduced size SGMV, but with a longer domain size. The springs and dimples of the spacer grid are not represented for the simplicity in the present investigation. The mean flow velocity and turbulent kinetic energy profiles are averaged from the instantaneous DNS data in different types of geometries as well as various downstream locations with respect to SGMV region. The results will help better assess the influence of spacer grid and mixing vanes on core coolant flows.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

On-demand data analytics in HPC environments at leadership computing facilities: Challenges and experiences

Scaling-up scientific data analysis and machine learning algorithms for data-driven discovery is a grand challenge that we face today. Despite the growing need for analysis from science domains that are generating `Big Data' from instruments and simulations, building high-performance analytical workflows of data-intensive algorithms have been daunting because: (i) the `Big Data' hardware and software architecture landscape is constantly evolving, (ii) newer architectures impose new programming models, and (iii) data-parallel kernels of analysis algorithms and their performance facets on different architectures are poorly understood. To address these problems, we have: (i) identified scalable data-parallel kernels of popular data analysis algorithms, (ii) implemented `Mini-Apps' of those kernels using different programming models (e.g. Map Reduce, MPI, etc.), (iii) benchmarked and validated the performance of the kernels in diverse architectures. In this paper, we discuss two of those Mini-Apps and show the execution of principal component analysis built as a workflow of the Mini-Apps. We show that Mini-Apps enable scientists to (i) write domain-specific data analysis code that scales on most HPC hardware and (ii) and offers the ability (most times with over a 10x speed-up) to analyze data sizes 100 times the size of what off-the-shelf desktop/workstations of today can handle.

Harney, John↗

Bayesian Abel Inversion in Quantitative X-Ray Radiography

A common image formation process in high-energy X-ray radiography is to have a pulsed power source that emits X-rays through a scene, a scintillator that absorbs X-rays and uoresces in the visible spectrum in response to the absorbed photons, and a CCD camera that images the visible light emitted from the scintillator. The intensity image is related to areal density, and, for an object that is radially symmetric about a central axis, the Abel transform then gives the object's volumetric density. Two of the primary drawbacks to classical variational methods for Abel inversion are their sensitivity to the type and scale of regularization chosen and the lack of natural methods for quantifying the uncertainties associated with the reconstructions. In this work we cast the Abel inversion problem within a statistical framework in order to compute volumetric object densities from X-ray radiographs and to quantify uncertainties in the reconstruction. A hierarchical Bayesian model is developed with a likelihood based on a Gaussian noise model and with priors placed on the unknown density pro le, the data precision matrix, and two scale parameters. This allows the data to drive the localization of features in the reconstruction and results in a joint posterior distribution for the unknown density pro le, the prior parameters, and the spatial structure of the precision matrix. Results of the density reconstructions and pointwise uncertainty estimates are presented for both synthetic signals and real data from a U.S. Department of Energy X-ray imaging facility.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗