Search NASA⌕ Search

SEARCH · Search NASA

Results for “accelerated computing”

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 847 records · Page 47

Why Is Improvement of Earth System Models so Elusive? Challenges and Strategies from Dust Aerosol Modeling

Past decades have seen an accelerating increase in computing efficiency, while climate models are representing a rapidly widening set of physical processes. Yet simulations of some fundamental aspects of climate like precipitation or aerosol forcing remain highly uncertain and resistant to progress. Dust aerosol modeling of soil particles lofted by wind erosion has seen a similar conflict between increasing model sophistication and remaining uncertainty. Dust aerosols perturb the energy and water cycles by scattering radiation and acting as ice nuclei, while mediating atmospheric chemistry and marine photosynthesis (and thus the carbon cycle). These effects take place across scales from the dimensions of an ice crystal to the planetary-scale circulation that disperses dust far downwind of its parent soil. Representing this range leads to several modeling challenges. Should we limit complexity in our model, which consumes computer resources and inhibits interpretation? How do we decide if a process involving dust is worthy of inclusion within our model? Can we identify a minimal representation of a complex process that is efficient yet retains the physics relevant to climate? Answering these questions about the appropriate degree of representation is guided by model evaluation, which presents several more challenges. How do we proceed if the available observations do not directly constrain our process of interest? (This could result from competing processes that influence the observed variable and obscure the signature of our process of interest.) Examples will be presented from dust modeling, with lessons that might be more broadly applicable. The end result will either be clinical depression or there assuring promise of continued gainful employment as the community confronts these challenges.

Climate models↗

NASA Advanced Supercomputing Facility Expansion

The NASA Advanced Supercomputing (NAS) Division enables advances in high-end computing technologies and in modeling and simulation methods to tackle some of the toughest science and engineering challenges facing NASA today. The name "NAS" has long been associated with leadership and innovation throughout the high-end computing (HEC) community. We play a significant role in shaping HEC standards and paradigms, and provide leadership in the areas of large-scale InfiniBand fabrics, Lustre open-source filesystems, and hyperwall technologies. We provide an integrated high-end computing environment to accelerate NASA missions and make revolutionary advances in science. Pleiades, a petaflop-scale supercomputer, is used by scientists throughout the U.S. to support NASA missions, and is ranked among the most powerful systems in the world. One of our key focus areas is in modeling and simulation to support NASA's real-world engineering applications and make fundamental advances in modeling and simulation methods.

NASA↗

Highly Polarized Ion Sources for Electron Ion Colliders (EIC)

The operation of the RHIC facility at BNL and the Electron Ion Colliders (EIC) under development at Jefferson Laboratory and BNL need high brightness ion beams with the highest polarization. Charge exchange injection into a storage ring or synchrotron and Siberian snakes have the potential to handle the needed polarized beam currents, but first the ion sources must create beams with the highest possible polarization to maximize collider productivity, which is proportional to a high power of the polarization. We are developing one universal H-/D- ion source design which will synthesize the most advanced developments in the field of polarized ion sources to provide high current, high brightness, ion beams with greater than 90% polarization, good lifetime, high reliability, and good power efficiency. The new source will be an advanced version of an atomic beam polarized ion source (ABPIS) with resonant charge exchange ionization by negative ions. An integrated ABPIS design will be prepared based on new materials and an optimized magnetic focusing system. Polarized atomic and ion beam formation, extraction, and transport for the new source will be computer simulated.

43 PARTICLE ACCELERATORS↗

Prototyping and Development of Commercial Nanocrystalline and Thin Film Silicon for Photovoltaic Manufacturing

The College of Nanoscale Science and Engineering (CNSE) at the University at Albany received funding from the Department of Energy for its proposal Prototyping and Development of Commercial Nanocrystalline and Thin Film Si for Photovoltaic Manufacturing. This project was created to identify growth rate, texture uniformity, process window, economics, composition and thickness uniformity solutions related to fabricating large area, high efficiency thin film silicon based solar cells. This document serves as a final report for the closure of this program and details the deliverables from CNSE against its original scope of work. Thin-film silicon solar cells are a promising candidate for electricity generation applications because of a combination of advantages. Nanocrystalline and poly-Si based thin films, reduces the use of expensive semiconductor material content, can be deposited onto a foreign substrate (e.g. glass or flexible stainless steel) and enables use of the cells in wide variety of applications. In addition, nano and poly-Si films have higher carrier mobility as well as reduce recombination effects, relative to traditional amorphous-silicon films. They can be mass-produced at low cost, and expected to have a strong position in the international photovoltaic industry, which is experiencing a compounded annual growth of 25%. The objectives included: Demonstration of high rate VHF (Very High Frequency) growth of nc-Si over large areas with uniform thickness. Demonstration of single chamber device growth that allows mass production processing. Demonstration of uniform segmented electrodes. Development of computer models to accelerate efforts. Demonstration of large grain thin film polycrystalline silicon films fabrication. Utilizing the AIC (Aluminum Induced Crystallization) process for large grain silicon film fabrication. Demonstration of 12% efficient nanocrystalline and polycrystalline thin film devices. The end result expected was the production of high-efficiency, low-cost thin film Si devices with reduced cost per watt, allowing dramatically increased market penetration. While further research is required to meet these ultimate goals, this project proved its feasibility.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Multiscale/Multiphysics Modeling of Biomass Thermochemical Processes

Computational problems in simulating biomass thermochemical processes involve coupling processes that span several orders of magnitude in space and time. Computational difficulties arise from the multitude of the problem governing equations, each typically applying over a narrow range of spatiotemporal scales, thus making it necessary to represent the processes as the result of the interaction of multiple physics modules, termed here as multiscale/multiphysics (MSMP) coupling. Predictive simulations for such processes require algorithms that can efficiently integrate the underlying MSMP methods across the scales in order to achieve prescribed accuracy and control the computational cost. In addition, MSMP algorithms must scale to one hundred thousand processors or more in order to effectively harness the new computational resources and accelerate the scientific advances. In this chapter, we discuss the state-of-the-art in modeling the macro-scale phenomena in a biomass pyrolysis reactor along with details of the shortcomings and prospects in improving predictability. We also introduce the various multiphysics modules needed to model thermochemical conversion at lower spatiotemporal scales. Furthermore, we illustrate the need for MSMP coupling for thermochemical processes in biomass and provide an overview of the wavelet-based coupling techniques we have developed recently. In particular, we provide details about the compound wavelet matrix (CWM) and the dynamic CWM (dCWM) methods and show they are highly efficient in transferring information among multiphysics models across multiple temporal and spatial scales. The algorithmic gain is in addition to the parallel spatial scalability from traditional domain decomposition methods. The CWM algorithms are serial in time and limited by the smallest-system time-scales. In order to relax this algorithmic constraint, we have recently coupled time parallel (TP) algorithms to CWM, thus yielding a novel approach termed tpCWM. We present preliminary results from the tpCWM technique, indicating that we can accelerate time-to-solution by 2 to 3-orders of magnitude even on 20-processors and this can potentially constitute a new paradigm for MSMP simulations. If such improvements in simulation capability can be generalized, the tpCWM approach can lead the way to predictive simulations of biomass thermochemical processes.

09 BIOMASS FUELS↗

Experiences with High-Level Programming Directives for Porting Applications to GPUs

HPC systems now exploit GPUs within their compute nodes to accelerate program performance. As a result, high-end application development has become extremely complex at the node level. In addition to restructuring the node code to exploit the cores and specialized devices, the programmer may need to choose a programming model such as OpenMP or CPU threads in conjunction with an accelerator programming model to share and manage the difference node resources. This comes at a time when programmer productivity and the ability to produce portable code has been recognized as a major concern. In order to offset the high development cost of creating CUDA or OpenCL kernels, directives have been proposed for programming accelerator devices, but their implications are not well known. In this paper, we evaluate the state of the art accelerator directives to program several applications kernels, explore transformations to achieve good performance, and examine the expressiveness and performance penalty of using high-level directives versus CUDA. We also compare our results to OpenMP implementations to understand the benefits of running the kernels in the accelerator versus CPU cores.

43 PARTICLE ACCELERATORS↗

The Science Training Program for Young Italian Physicists and Engineers at Fermilab

Since 1984 Fermilab has been hosting a two-month summer training program for selected undergraduate and graduate Italian students in physics and engineering. Building on the traditional close collaboration between the Italian National Institute of Nuclear Physics (INFN) and Fermilab, the program is supported by INFN, by the DOE and by the Scuola Superiore di Sant`Anna of Pisa (SSSA), and is run by the Cultural Association of Italians at Fermilab (CAIF). This year the University of Pisa has qualified it as a “University of Pisa Summer School”, and will grant successful students with European Supplementary Credits. Physics students join the Fermilab HEP research groups, while engineers join the Particle Physics, Accelerator, Technical, and Computing Divisions. Some students have also been sent to other U.S. laboratories and universities for special trainings. The programs cover topics of great interest for science and for social applications in general, like advanced computing, distributed data analysis, nanoelectronics, particle detectors for earth and space experiments, high precision mechanics, applied superconductivity. In the years, over 350 students have been trained and are now employed in the most diverse fields in Italy, Europe, and the U.S. In addition, the existing Laurea Program in Fermilab Technical Division was extended to the whole laboratory, with presently two students in Master’s thesis programs on neutrino physics and detectors in the Neutrino Division. And finally, a joint venture with the Italian Scientists and Scholars North-America Foundation (ISSNAF) provided this year 4 professional engineers free of charge for Fermilab. More details on all of the above can be found below.

Barzi, Emanuela↗

Livermore Big Artificial Neural Network Toolkit

LBANN is a toolkit that is designed to train artificial neural networks efficiently on high performance computing architectures. It is optimized to take advantages of key High Performance Computing features to accelerate neural network training. Specifically it is optimized for low-latency, high bandwidth interconnects, node-local NVRAM, node-local GPU accelerators, and high bandwidth parallel file systems. It is built on top of the open source Elemental distributed-memory dense and spars-direct linear algebra and optimization library that is released under the BSD license. The algorithms contained within LBANN are drawn from the academic literature and implemented to work within a distributed-memory framework.

Essen, Brian Van↗

Dimensionality Reduction Library v 0.2

Dimensionality Reduction Library is a C++ library for dimensionality reduction. In the context of this library, dimensionality reduction is considered to consist of 1)estimation of the intrinsic dimensionality using sampled data, 2) Finding maps that reduce the diemsionality of data (forward map) or increase the dimensionality of data (reverse map) and 3) mapping arbitray coordiantes to high and low dimensionalities. The library is intended toprovide a consistent interface to multiple dimensionality reduction algorithms with an efficient C++ interface that runs efficiently on multicore architectures. A few routines have been optimized with an option for GPU acceleration or distributed computation. Currently the library offers intrinsic dimensionality estimation using point-PCA reconstruction error and/ residual variance. The following dimensionality reduction methods have been implemented: Principal Component Analysis Multidimensional Scaling Locally Linear Embedding IsoMap Autoencoder Neutral Networks An executable is also supplied that can be built to allow for command-line access to the library routines. A description for an applciation of the library for molecular structure analysis has been published.

BROWN, WILLIAM↗

The Fermilab Connection

More than 4,000 scientists in 53 countries use Fermilab and its particle accelerators, detectors and computers for their research. That includes about 2,500 scientists from 223 U.S. institutions in 42 states, plus the District of Columbia and Puerto Rico.

Fermilab↗

Studying Performance Portability of LAMMPS across Diverse GPU-based Platforms

The molecular dynamics simulation software, LAMMPS, utilizes the Kokkos acceleration library to port computation to a diverse set of architectures including those based on GPU accelerators. In addition to Kokkos, LAMMPS contains a vast code base that leverages the CUDA application programming interface using library functions such as cuFFT, CUDA’s fast-fourier transform (FFT) library, and, more recently, also support for AMD’s Heterogeneous Interface for Portability (HIP) that is rapidly growing. While preparing LAMMPS tests for the AMD GPU-based test system precursors to Frontier, we investigated several strategies for accelerating LAMMPS on AMD GPUs, using the AMD Instinct MI100 and MI250X. In this work, we integrated the HIP FFT library, hipFFT, into the particle-particle particle-mesh (PPPM) long-range solver, which allowed the porting of PPPM calculations to the GPUs. Kokkos behavior on the MI100 and MI250X was also investigated through the package kokkos command of LAMMPS, targeting com- munication, memory usage, and particle grid decomposition. The Tersoff, Reax, Lennard-Jones (LJ), EAM, Granular, and PPPM potentials were investigated in this effort, and results from these experiments are provided. The selected potentials were run on Spock (AMD Instinct MI100), Crusher (AMD Instinct MI250X), AFW HPC11 (NVIDIA A100) and Summit (NVIDIA V100), for comparison. Operational roofline models were constructed and analyzed for the Tersoff, Reax, and Lennard-Jones potentials on Crusher and Summit.

Hagerty, Nick↗

Modification on Surface Oxide Layer Structure and Surface Morphology of Niobium by GCIB Treatments

Recently, it was demonstrated1 that significant reductions in field emission on Nb surfaces could be achieved by means of a new surface treatment technique called gas cluster ion beam (GCIB). Further study revealed that GCIB treatments could also modify surface irregularities and remove surface asperities leading to a smoother surface finish as demonstrated through measurements using a 3-D profilometer, an atomic force microscope, and a scanning electron microscope. These experimental observations were supported by computer simulation via atomistic molecular dynamics and a phenomenological surface dynamics. Measurements employing a secondary ion mass spectrometry found that GCIB could also alter Nb surface oxide layer structure. Possible implications of the experimental results on the performance of Nb superconducting radio frequency cavities treated by GCIB will be discussed. 1: A.T. Wu et al, invited contribution to the book ?Neural Computation and Particle Accelerators: Research, Technology, and Applications?, NOVA Science Publishers, New York, USA, 2009, In press

Wu, Andy↗

Acceleration and torque feedback for robotic control - Experimental results

Gross motion control of robotic manipulators typically requires significant on-line computations to compensate for nonlinear dynamics due to gravity, Coriolis, centripetal, and friction nonlinearities. One controller proposed by Luo and Saridis avoids these computations by feeding back joint acceleration and torque. This study implements the controller on a Puma 600 robotic manipulator. Joint acceleration measurement is obtained by measuring linear accelerations of each joint, and deriving a computationally efficient transformation from the linear measurements to the angular accelerations. Torque feedback is obtained by using the previous torque sent to the joints. The implementation has stability problems on the Puma 600 due to the extremely high gains inherent in the feedback structure. Since these high gains excite frequency modes in the Puma 600, the algorithm is modified to decrease the gain inherent in the feedback structure. The resulting compensator is stable and insensitive to high frequency unmodeled dynamics. Moreover, a second compensator is proposed which uses acceleration and torque feedback, but still allows nonlinear terms to be fed forward. Thus, by feeding the increment in the easily calculated gravity terms forward, improved responses are obtained. Both proposed compensators are implemented, and the real time results are compared to those obtained with the computed torque algorithm.

Mclnroy, John E.↗

Accelerating kinetic plasma simulations with machine-learning-generated initial conditions

Computational models of plasma technologies often solve for the system operating conditions by time-stepping an initial value problem to a quasi-steady solution. However, the strongly nonlinear and multi-timescale nature of plasma dynamics often necessitate millions, or even hundreds of millions, of steps to reach convergence, reducing the effectiveness of these simulations for computer-aided engineering. We consider acceleration of kinetic plasma simulations via data-driven machine-learning-generated initial conditions, which initialize the simulations close to their final quasi-steady-state, thereby reducing the number of steps to reach convergence. Three machine-learning models are developed to predict the density and ion kinetic profiles of capacitively coupled plasma discharges relevant to the microelectronics industry. The models are trained on kinetic simulations over a range of device operating frequencies and pressures. Best performance was observed when simulations were initialized with ion kinetic profiles generated by a convolutional neural network, reducing the mean number of steps to reach convergence by 17.1× when compared to initialization with a zero-dimensional global model. We also outline a workflow for continuous data-driven model improvement and simulation speedup, with the aim of generating sufficient data for full device digital twins.

Artificial neural networks↗

Process Simulation Role in the Development of New Alloys Based on Integrated Computational Material Science and Engineering

To accelerate the introduction of new materials and components, the development of metal casting processes requires the teaming between different disciplines, as multi-physical phenomena have to be considered simultaneously for the process design and optimization of mechanical properties. The required models for physical phenomena as well as their validation status for metal casting are reviewed. The data on materials properties, model validation, and relevant microstructure for materials properties are highlighted. One vehicle to accelerate the development of new materials is through combined experimental-computational efforts. Integrated computational/experimental practices are reviewed; strengths and weaknesses are identified with respect to metal casting processes. Specifically, the examples are given for the knowledge base established at Oak Ridge National Laboratory and computer models for predicting casting defects and microstructure distribution in aluminum alloy components.

Sabau, Adrian S↗

Neural network methods for radiation detectors and imaging

Recent advances in image data proccesing through deep learning allow for new optimization and performance-enhancement schemes for radiation detectors and imaging hardware. This enables radiation experiments, which includes photon sciences in synchrotron and X-ray free electron lasers as a subclass, through data-endowed artificial intelligence. We give an overview of data generation at photon sources, deep learning-based methods for image processing tasks, and hardware solutions for deep learning acceleration. Most existing deep learning approaches are trained offline, typically using large amounts of computational resources. However, once trained, DNNs can achieve fast inference speeds and can be deployed to edge devices. A new trend is edge computing with less energy consumption (hundreds of watts or less) and real-time analysis potential. While popularly used for edge computing, electronic-based hardware accelerators ranging from general purpose processors such as central processing units (CPUs) to application-specific integrated circuits (ASICs) are constantly reaching performance limits in latency, energy consumption, and other physical constraints. These limits give rise to next-generation analog neuromorhpic hardware platforms, such as optical neural networks (ONNs), for high parallel, low latency, and low energy computing to boost deep learning acceleration (LA-UR-23-32395).

edge computing↗

Quantum-centric supercomputing for materials science: A perspective on challenges and future directions

Computational models are an essential tool for the design, characterization, and discovery of novel materials. Computationally hard tasks in materials science stretch the limits of existing high-performance supercomputing centers, consuming much of their resources for simulation, analysis, and data processing. Quantum computing, on the other hand, is an emerging technology with the potential to accelerate many of the computational tasks needed for materials science. In order to do that, the quantum technology must interact with conventional high-performance computing in several ways: approximate results validation, identification of hard problems, and synergies in quantum-centric supercomputing. Here in this paper, we provide a perspective on how quantum-centric supercomputing can help address critical computational problems in materials science, the challenges to face in order to solve representative use cases, and new suggested directions.

36 MATERIALS SCIENCE↗

Quantum-centric Supercomputing for Materials Science: A Perspective on Challenges and Future Directions

Computational models are an essential tool for the design, characterization, and discovery of novel materials. Hard computational tasks in materials science stretch the limits of existing high-performance supercomputing centers, consuming much of their simulation, analysis, and data resources. Quantum computing, on the other hand, is an emerging technology with the potential to accelerate many of the computational tasks needed for materials science. In order to do that, the quantum technology must interact with conventional high-performance computing in several ways: approximate results validation, identification of hard problems, and synergies in quantum-centric supercomputing. In this paper, we provide a perspective on how quantum-centric supercomputing can help address critical computational problems in materials science, the challenges to face in order to solve representative use cases, and new suggested directions.

36 MATERIALS SCIENCE↗