Search NASA⌕ Search

SEARCH · Search NASA

Results for “High efficiency acceleration”

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 595 records · Page 33
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

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↗
DOE OSTI2023

A high-performance lung-inspired ceramic 3D-printed heat exchanger for high-temperature energy-efficient systems

The thermal-to-electric conversion efficiency of next-generation supercritical carbon dioxide and concentrated solar power plants is significantly augmented by operating at temperatures exceeding 1000 °C if durable, compact, high-temperature heat exchangers (HXs) are developed. Whereas metal 3D-printed HXs fail to operate under extreme temperature conditions, ceramic 3D-printed HXs are deemed promising candidates for high-temperature, highly oxidizing environments. Ceramic 3D-printed HXs, however, exhibit a high gas permeability issue through hot and cold separating walls due to an excessive non-volatile photopolymer content utilized. Here, in this work, a novel leak-free lung-inspired ceramic 3D-printed HX employing a highly complex heat transfer topology is introduced for high-temperature energy-efficient systems. The high permeability issue of ceramic 3D-printed HXs is fully eliminated through a uniform zinc-based coating. It was determined that the effective thermal conductivity of the sintered ceramic heat exchangers is affected by the coated layer. The high-temperature thermal and hydraulic characteristics of the lung-inspired ceramic 3D-printed HX are comprehensively examined and compared against a millichannel 3D-printed HX module. Experimental results indicated that the lung-inspired 3D-printed ceramic HX significantly outperforms its millichannel counterpart. The lung-inspired HX shows a volume-based power density of 8.2 MW/m3 at a hot-side inlet air temperature of 700°C, a 71 % improvement compared with that of the millichannel HX. More importantly, the higher thermal duty of the lung-inspired 3D-printed HX is realized at a lower normalized pressure drop penalty of 510 Pa/W, 22 % lower than that of the millichannel ceramic HX. This study accelerates the advancement of high-performance ceramic 3D-printed HXs with complex topologies for high-temperature energy-efficient systems and/or corrosive environments.

42 ENGINEERING↗
NASA NTRS2025

An Efficient GPU-Accelerated Multi-Source Global Fit Pipeline for LISA Data Analysis

The large-scale analysis task of deciphering gravitational wave signals in the LISA data stream will be difficult, requiring a large amount of computational resources and extensive development of computational methods. Its high dimensionality, multiple model types, and complicated noise profile require a global fit to all parameters and input models simultaneously. In this work, we detail our global fit algorithm, called “Erebor,” designed to accomplish this challenging task. It is capable of analysing current state-of-the-art datasets and then growing into the future as more pieces of the pipeline are completed and added. We describe our pipeline strategy, the algorithmic setup, and the results from our analysis of the LDC2A Sangria dataset, which contains Massive Black Hole Binaries, compact Galactic Binaries, and a parameterized noise spectrum whose parameters are unknown to the user. The Erebor algorithm includes three unique and very useful contributions: GPU acceleration for enhanced computational efficiency; ensemble MCMC sampling with multiple MCMC walkers per temperature for better mixing and parallelized sample creation; and special online updates to reversible-jump (or trans-dimensional) sampling distributions to ensure sampler mixing and accurate initial estimates for detectable sources in the data. We recover posterior distributions for all 15 (6) of the injected MBHBs in the LDC2A training (hidden) dataset. We catalog ∼12000 Galactic Binaries (∼8000 as high confidence detections) for both the training and hidden datasets. All of the sources and their posterior distributions are provided in publicly available catalogs.

LISA global fit↗
DOE OSTI2025

Toward Fullerene-Free PIN Perovskite Solar Cells

We highlight opportunities for a transformative shift in perovskite solar cell design by expanding electron transport layers (ETLs) beyond fullerenes. Fullerenes have known limitations, including constraints on open-circuit voltage, stability, and mechanical integrity. Recently, fullerene-free p-i-n cells with power conversion efficiencies exceeding 25% have been demonstrated via both naphthalene diimide-SnO x bilayers and nonfullerene acceptor-based ETLs. Despite successes, fullerenes remain the de facto ETLs for perovskites. Drawing lessons from organic photovoltaics, where it took decades to transition from fullerenes to more broadly available and efficient materials, we explore pathways to accelerate the development and adoption of fullerene-free ETLs. This requires understanding the similarities and differences between organic and perovskite solar cells, which will necessitate carefully designing fullerene replacements with both, high efficiency and also, critically, durability under operation. Here, we incorporate literature data to facilitate comparisons, and independently conduct fracture energy measurements for alternative ETL configurations to motivate their adoption.

14 SOLAR ENERGY↗
NASA NTRS2025

Efficient GPU-Accelerated MultiSource Global Fit Pipeline for LISA Data Analysis

The large-scale analysis task of deciphering gravitational-wave signals in the LISA data stream will be difficult, requiring a large amount of computational resources and extensive development of computational methods. Its high dimensionality, multiple model types, and complicated noise profile require a global fit to all parameters and input models simultaneously. In this work, we detail our global fit algorithm, called “Erebor,” designed to accomplish this challenging task. It is capable of analyzing current state-of-the-art datasets and then growing into the future as more pieces of the pipeline are completed and added. We describe our pipeline strategy, the algorithmic setup, and the results from our analysis of the LDC2A Sangria dataset, which contains massive black hole binaries, compact galactic binaries, and a parametrized noise spectrum whose parameters are unknown to the user. The Erebor algorithm includes three unique and very useful contributions: GPU acceleration for enhanced computational efficiency; ensemble Markov Chain Monte Carlo (MCMC) sampling with multiple MCMC walkers per temperature for better mixing and parallelized sample creation; and special online updates to reversible-jump (or transdimensional) sampling distributions to ensure sampler mixing and accurate initial estimates for detectable sources in the data.We recover posterior distributions for all 15 (6) of the injected massive black hole binaries (MBHB) in the LDC2A training (hidden) dataset. We catalog ∼12000 galactic binaries (∼8000 as high confidence detections) for both the training and hidden datasets. All of the sources and their posterior distributions are provided in publicly available catalogs.

LISA↗
DOE OSTI2025

Even Higher-Level Synthesis: An Exploration of AI Hardware Accelerators using HLS4ML

With the rise of artificial intelligence, the popularization of deep learning, and a constantly evolving industry, the demand for flexible and efficient tools has never been greater. As algorithms grow more complex, their runtime and energy consumption increase exponentially. Customized hardware accelerators, long used for specific mathematical operations, remain essential for managing modern applications' computational and power demands. Hardware accelerators can speed up complex computations by orders of magnitude, but their manual design and verification processes are often challenging and time-consuming. High-Level Synthesis (HLS) provides a solution by transforming high-level algorithm descriptions, typically written in C++ or SystemC, into synthesizable RTL suitable for hardware implementation. This approach reduces development time for RTL engineers while offering flexibility beyond what traditional handwritten RTL can provide. We extended this capability to the machine-learning domain with the open-source framework hls4ml, which allows neural networks trained in Python frameworks like Tensorflow or PyTorch to be synthesized into efficient hardware representations for the traditional FPGA and ASIC flows. This breakthrough addresses the growing need for reduced design turnaround and easy verification of ML hardware accelerators with low latency and power efficiency constraints. During this tutorial, we will demonstrate how Python complements HLS by simplifying the ML design process, bridging the gap between software and hardware development. Attendees will explore how we translate neural networks modeled in Python into fixed-point C++ models suitable for HLS workflows. We will dive into strategies like Value-Range Analysis and Quantization-Aware Training, which optimize these designs for deployment and evaluate their accuracy, power consumption, and energy efficiency. To exemplify these concepts, experts from Fermilab will share their experiences applying this technology to high-energy physics experiments, where real-time, low-latency processing is critical. Over the years, Fermilab engineers have demonstrated how deep neural networks, optimized for hardware using hls4ml, can meet the stringent requirements of trigger systems at the CERN Large Hadron Collider. These systems rely on rapid decision-making to process immense data volumes while retaining only the most relevant events for further analysis. The application of hls4ml has also been extended to innovative technologies like smart pixel arrays. These smart pixels integrate ML inference capabilities directly into sensor devices, enabling localized data processing at the pixel level. This approach drastically reduces the need to transmit raw data to external processing units, significantly decreasing power consumption and latency. By embedding neural networks within the pixel architecture, the smart pixels can identify and prioritize relevant data in real time, providing a highly efficient solution for edge computing in scenarios such as particle detectors and imaging systems. Fermilab's work highlights the potential of hardware-accelerated ML in scenarios where both speed and power efficiency are mission-critical. Through this tutorial, attendees will gain valuable insights into the challenges and solutions of deploying ML in hardware. Understanding how HLS and hls4ml streamline the development of neural network-based hardware accelerators is fundamental for the industry's future. Participants will learn how these technologies are shaping the future of AI and scientific computing.

Di Guglielmo, Giuseppe [Fermilab]↗
DOE OSTI2017

Laser-plasmas in the relativistic-transparency regime: science and applications

Laser-plasma interactions in the novel regime of relativistically induced transparency (RIT) have been harnessed to generate intense ion beams efficiently with average energies exceeding 10 MeV/nucleon (>100 MeV for protons) at “table-top” scales in experiments at the LANL Trident Laser. By further optimization of the laser and target, the RIT regime has been extended into a self-organized plasma mode. This mode yields an ion beam with much narrower energy spread while maintaining high ion energy and conversion efficiency. This mode involves self-generation of persistent high magnetic fields (~10 4 T, according to particle-in-cell simulations of the experiments) at the rear-side of the plasma. These magnetic fields trap the laser-heated multi-MeV electrons, which generate a high localized electrostatic field (~0.1 TV/m). After the laser exits the plasma, this electric field acts on a highly structured ion-beam distribution in phase space to reduce the energy spread, thus separating acceleration and energy-spread reduction. Thus, ion beams with narrow energy peaks at up to 18 MeV/nucleon are generated reproducibly with high efficiency (≈5%). The experimental demonstration has been done with 0.12 PW, high-contrast, 0.6 ps Gaussian 1.053 μm laser pulses irradiating planar foils up to 250 nm thick at 2–8 × 10 20 W/cm 2 . These ion beams with co-propagating electrons have been used on Trident for uniform volumetric isochoric heating to generate and study warm-dense matter at high densities. These beam plasmas have been directed also at a thick Ta disk to generate a directed, intense point-like Bremsstrahlung source of photons peaked at ~2 MeV and used it for point projection radiography of thick high density objects. In addition, prior work on the intense neutron beam driven by an intense deuterium beam generated in the RIT regime has been extended. Neutron spectral control by means of a flexible converter-disk design has been demonstrated, and the neutron beam has been used for point-projection imaging of thick objects. Finally, we discuss the plans and prospects for further improvements and applications.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2018

Cascade Transformer Based on the Volume Coil for Power Transmission under High Voltage

The use of an electronic cooling system at the High Intensity heavy ion Accelerator Facility (HIAF) accelerator complex, which is being developed at the Institute of Modern Physics (China), to improve the efficiency of ion injection into the accelerator and reduce the spread of ion pulses in the beam has been proposed. Electron cooling of the ion beam was carried out due to the interaction of ions with a continuous electron beam with a current of up to 3 A, energy of up to 450 keV, and energy stability at the level of 10{sup –4} or better. The electron beam energy recuperation was carried out at the expense of a power source with a power of 5–15 kW, which was located at the top of a high-voltage column—a high-voltage terminal. The operation of a prototype of power transmission system, which was based on a cascade transformer with a volumetric coil, has been considered. Such a transformer has a relatively low scattering inductance, which can significantly reduce the number of capacitors to compensate for it. It has been shown that this design made it possible to transfer power of up to 40 kW at small dimensions of the transformer and heat dissipation in it was not more than 10 kW.

43 PARTICLE ACCELERATORS↗
NASA NTRS1993

On the physics of waves in the solar atmosphere: Wave heating and wind acceleration

This paper presents work performed on the generation and physics of acoustic waves in the solar atmosphere. The investigators have incorporated spatial and temporal turbulent energy spectra in a newly corrected version of the Lighthill-Stein theory of acoustic wave generation in order to calculate the acoustic wave energy fluxes generated in the solar convective zone. The investigators have also revised and improved the treatment of the generation of magnetic flux tube waves, which can carry energy along the tubes far away from the region of their origin, and have calculated the tube energy fluxes for the sun. They also examine the transfer of the wave energy originated in the solar convective zone to the outer atmospheric layers through computation of wave propagation and dissipation in highly nonhomogeneous solar atmosphere. These waves may efficiently heat the solar atmosphere and the heating will be especially significant in the chromospheric network. It is also shown that the role played by Alfven waves in solar wind acceleration and coronal hole heating is dominant. The second part of the project concerned investigation of wave propagation in highly inhomogeneous stellar atmospheres using an approach based on an analytic tool developed by Musielak, Fontenla, and Moore. In addition, a new technique based on Dirac equations has been developed to investigate coupling between different MHD waves propagating in stratified stellar atmospheres.

Musielak, Z. E.↗
DOE OSTI2018

Design-to-Device Approach Affords Panchromatic Co-sensitized Solar Cells

Data–driven materials discovery has become increasingly important in identifying materials that exhibit specific, desirable properties from a vast chemical search space. Synergic prediction and experimental validation are needed to accelerate scientific advances related to critical societal applications. A design–to–device study that uses high–throughput screens with algorithmic encodings of structure–property relationships is reported to identify new materials with panchromatic optical absorption, whose photovoltaic device applications are then experimentally verified. The data–mining methods source 9431 dye candidates, which are auto–generated from the literature using a custom text–mining tool. These candidates are sifted via a data–mining workflow that is tailored to identify optimal combinations of organic dyes that have complementary optical absorption properties such that they can harvest all available sunlight when acting as co–sensitizers for dye–sensitized solar cells (DSSCs). Six promising dye combinations are shortlisted for device testing, whereupon one dye combination yields co–sensitized DSSCs with power conversion efficiencies comparable to those of the high–performance, organometallic dye, N719. Here, these results demonstrate how data–driven molecular engineering can accelerate materials discovery for panchromatic photovoltaic or other applications.

14 SOLAR ENERGY↗
NASA NTRS1994

On the physics of waves in the solar atmosphere: Wave heating and wind acceleration

This paper presents work performed on the generation and physics of acoustic waves in the solar atmosphere. The investigators have incorporated spatial and temporal turbulent energy spectra in a newly corrected version of the Lighthill-Stein theory of acoustic wave generation in order to calculate the acoustic wave energy fluxes generated in the solar convective zone. The investigators have also revised and improved the treatment of the generation of magnetic flux tube waves, which can carry energy along the tubes far away from the region of their origin, and have calculated the tube wave energy fluxes for the sun. They also examine the transfer of the wave energy originated in the solar convective zone to the outer atmospheric layers through computation of wave propagation and dissipation in highly nonhomogeneous solar atmosphere. These waves may efficiently heat the solar atmosphere and the heating will be especially significant in the chromospheric network. It is also shown that the role played by Alfven waves in solar wind acceleration and coronal hole heating is dominant. The second part of the project concerned investigation of wave propagation in highly inhomogeneous stellar atmospheres using an approach based on an analytic tool developed by Musielak, Fontenla, and Moore. In addition, a new technique based on Dirac equations has been developed to investigate coupling between different MHD waves propagating in stratified stellar atmospheres.

Musielak, Z. E.↗
DOE OSTI2022

GPU Acceleration of Large-Scale Full-Frequency GW Calculations

Many-body perturbation theory is a powerful method to simulate electronic excitations in molecules and materials starting from the output of density functional theory calculations. By implementing the theory efficiently so as to run at scale on the latest leadership high-performance computing systems it is possible to extend the scope of GW calculations. Here, we present a GPU acceleration study of the full-frequency GW method as implemented in the WEST code. Excellent performance is achieved through the use of (i) optimized GPU libraries, e.g., cuFFT and cuBLAS, (ii) a hierarchical parallelization strategy that minimizes CPU-CPU, CPU-GPU, and GPU-GPU data transfer operations, (iii) nonblocking MPI communications that overlap with GPU computations, and (iv) mixed precision in selected portions of the code. A series of performance benchmarks has been carried out on leadership high-performance computing systems, showing a substantial speedup of the GPU-accelerated version of WEST with respect to its CPU version. Good strong and weak scaling is demonstrated using up to 25 920 GPUs. Finally, we showcase the capability of the GPU version of WEST for large-scale, full-frequency GW calculations of realistic systems, e.g., a nanostructure, an interface, and a defect, comprising up to 10 368 valence electrons.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

ODU Optimization of Nb3Sn supttering on SRF structures

High efficiency Superconducting radio-frequency (SRF) cavities are the key technology to cost-effective Continuous Wave (CW) SRF accelerators. Recent advances in Nb3Sn coatings on the inner surface of SRF cavities showed the potential for this material to significantly cut the capital and operating cost of SRF accelerators. The best results are presently achieved with Nb3Sn films coated by using the so-called vapor diffusion technique. Several alternative approaches have been and are being investigated such as electrochemical deposition, bronze routes, and sputtering. Among these techniques, magnetron sputtering is a versatile technique, which has been demonstrated to deposit Nb3Sn films on SRF cavities. The technique requires further development to compete with the results demonstrated on the cavities coated with the vapor diffusion techniques. The team to develop the novel technique includes a graduate student from Old Dominion University (ODU) and other experts. ODU Prof. Elsayed-Ali directs the graduate student to coat samples and cavities in the new coating system and analyze samples with different surface science techniques. Dr. Grigory Eremeev from Fermilab assists with 2.6 GHz cavities for coating, cavity characterization at cryogenic temperature, and result analysis. The research is funded by the Office of High Energy Physics at the US Department of Energy at both Fermilab and ODU

43 PARTICLE ACCELERATORS↗
DOE OSTI2016

Efficient production of high-energy nonthermal particles during magnetic reconnection in a magnetically dominated ion-electron plasma

Magnetic reconnection is a leading mechanism for dissipating magnetic energy and accelerating nonthermal particles in Poynting-flux-dominated flows. We investigate nonthermal particle acceleration during magnetic reconnection in a magnetically dominated ion–electron plasma using fully kinetic simulations. For an ion–electron plasma with a total magnetization of ${\sigma }_{0}={B}^{2}/(4\pi n({m}_{i}+{m}_{e}){c}^{2})$, the magnetization for each species is ${\sigma }_{i}\sim {\sigma }_{0}$ and ${\sigma }_{e}\sim ({m}_{i}/{m}_{e}){\sigma }_{0}$, respectively. We have studied the magnetically dominated regime by varying σe = 103–105 with initial ion and electron temperatures ${T}_{i}={T}_{e}=5-20{m}_{e}{c}^{2}$ and mass ratio ${m}_{i}/{m}_{e}=1-1836$. Our results demonstrate that reconnection quickly establishes power-law energy distributions for both electrons and ions within several (2–3) light-crossing times. For the cases with periodic boundary conditions, the power-law index is $1\lt s\lt 2$ for both electrons and ions. The hard spectra limit the power-law energies for electrons and ions to be ${\gamma }_{{be}}\sim {\sigma }_{e}$ and ${\gamma }_{{bi}}\sim {\sigma }_{i}$, respectively. The main acceleration mechanism is a Fermi-like acceleration through the drift motions of charged particles. When comparing the spectra for electrons and ions in momentum space, the spectral indices sp are identical as predicted in Fermi acceleration. We also find that the bulk flow can carry a significant amount of energy during the simulations. Finally, we discuss the implication of this study in the context of Poynting-flux dominated jets and pulsar winds, especially the applications for explaining nonthermal high-energy emissions.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Symplectic machine learning model for fast simulation of space-charge effects

Symplectic simulation of space-charge effects is crucial for the design and operation of high-intensity particle accelerators. Traditional methods for simulating these effects are often computationally expensive, resulting in significant overhead. In this work, we introduce a generative model based on a U-Net architecture within a generative adversarial network framework to efficiently simulate space-charge effects. The model is trained to predict the transverse multiparticle space-charge Hamiltonian, which can be physically computed using a gridless spectral method. The one-step symplectic transverse transfer map for the particles is then obtained by differentiating the predicted Hamiltonian. Benchmarking results demonstrate that this generative model achieves an order of magnitude higher computational efficiency compared to the spectral method, providing a highly efficient alternative for simulating space-charge effects with a large number of particles. By maintaining symplecticity, the model effectively preserves the phase-space structure and mitigates nonphysical errors in long-term simulations. This model has been integrated into jutrack, a novel autodifferentiable accelerator modeling code developed in the julia programming language.

Beam code development & simulation techniques↗
DOE OSTI2014

Directed evolution of an ultrastable carbonic anhydrase for highly efficient carbon capture from flue gas

Carbonic anhydrase (CA) is one of nature’s fastest enzymes and can dramatically improve the economics of carbon capture under demanding environments such as coal-fired power plants. The use of CA to accelerate carbon capture is limited by the enzyme’s sensitivity to the harsh process conditions. Using directed evolution, the properties of a β-class CA from Desulfovibrio vulgaris were dramatically enhanced. Iterative rounds of library design, library generation, and high-throughput screening identified highly stable CA variants that tolerate temperatures of up to 107 °C in the presence of 4.2 M alkaline amine solvent at pH >10.0. This increase in thermostability and alkali tolerance translates to a 4,000,000-fold improvement over the natural enzyme. In conclusion, at pilot scale, the evolved catalyst enhanced the rate of CO2 absorption 25-fold compared with the noncatalyzed reaction.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2023

Batched Sparse Linear Algebra (Final Report for Subcontract B648960)

This report finalizes design specifications for developing batched kernels for small tensor operations for unassembled matrix-free iterative solvers, batched solvers for partially assembled operators, and batched solvers with support for various sparse formats. The outcome of the project milestones is a set of interfaces to Batched Sparse LA solvers running on hardware accelerators for use in ECP Libraries and Applications. It is part of the development of sparse batched kernels, solvers/preconditioners as well as creating interoperability in xSDK libraries with sparse and dense batched functions to benefit ECP applications. The participants included representatives from ECP libraries (not limited to the xSDK project), applications, and vendors (AMD, Intel, and NVIDIA). Batched sparse linear algebra solvers form the new frontier for algorithmic development and performance engineering. Many applications (ECP and non-ECP alike) require simultaneous solutions of small linear systems of equations that are structurally sparse. To move towards high hardware utilization, it is important to provide these applications with appropriate interfaces to efficient batched sparse solvers running on modern hardware accelerators. We present interface designs in use by HPC software libraries supporting batched sparse linear algebra and the development of sparse batched kernel codes for solvers and preconditioners. We also address the potential interoperability opportunities to keep the software portable between the major hardware accelerators from AMD, Intel, and NVIDIA. The presented interface specifications includes batched band, sparse iterative, and sparse direct solvers. This report summarizes progress in Kokkos Kernels and the xSDK libraries MAGMA, Ginkgo, hypre, SUNDIALS, and SuperLU_dist.

97 MATHEMATICS AND COMPUTING↗