Search NASA⌕ Search

SEARCH · Search NASA

Results for “Accelerator Applications”

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 343 records · Page 19
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Toward a Holistic Performance Evaluation of Large Language Models Across Diverse AI Accelerators

Artificial intelligence (AI) methods have become critical in scientific applications to help accelerate scientific discovery. Large language models (LLMs) are being considered a promising approach to address some challenging problems because of their superior generalization capabilities across domains. The effectiveness of the models and the accuracy of the applications are contingent upon their efficient execution on the underlying hardware infrastructure. Specialized Al accelerator hardware systems have recently become available for accelerating Al applications. However, the comparative performance of these AI accelerators on large language models has not been previously studied. In this paper, we systematically study LLMs on multiple AI accelerators and GPUs and evaluate their performance characteristics for these models. We evaluate these systems with (i) a micro-benchmark using a core transformer block, (ii) a GPT-2 model, and (iii) an 1,I,M-driven science use case, GenSLM. We present our findings and analyses of the models' performance to better understand the intrinsic capabilities of AI accelerators. Furthermore, our analysis takes into account key factors such as sequence lengths, scaling behavior, and sensitivity to gradient accumulation steps.

Emani, Murali↗
DOE OSTI2022

Hvac: Removing I/O Bottleneck for Large-Scale Deep Learning Applications

Scientific communities are increasingly adopting deep learning (DL) models in their applications to accelerate scientific discovery processes. However, with rapid growth in the computing capabilities of HPC supercomputers, large-scale DL applications have to spend a significant portion of training time performing I/O to a parallel storage system. Previous research works have investigated optimization techniques such as prefetching and caching. Unfortunately, there exist non-trivial challenges to adopting the existing solutions on HPC supercomputers for large-scale DL training applications, which include non-performance and/or failures at extreme scale, lack of portability and generality in design, complex deployment methodology, and being limited to a specific application or dataset. To address these challenges, we propose High-Velocity AI Cache (HVAC), a distributed read-cache layer that targets and fully exploits the node-local storage or near node-local storage technology. HVAC seamlessly accelerates read I/O by aggregating node-local or near node-local storage, avoiding metadata lookups and file locking while preserving portability in the application code. We deploy and evaluate HVAC on 1,024 nodes (with over 6000 NVIDIA V100 GPUS) of the Summit supercomputer. In particular, we evaluate the scalability, efficiency, accuracy, and load distribution of HVAC compared to GPFS and XFS-on-NVMe. With four different DL applications, we observe an average 25 % performance improvement atop GPFS and 9% drop against XFS-on-NVMe, which scale linearly and are considered the performance upper bound. We envision HVAC as an important caching library for upcoming HPC supercomputers such as Frontier.

Khan, Awais↗
DOE OSTI2016

Modern compact accelerators of cyclotron type for medical applications

Ion beam therapy and hadron therapy are types of external beam radiotherapy. Recently, the vast majority of patients have been treated with protons and carbon ions. Typically, the types of accelerators used for therapy were cyclotrons and synchrocyclotrons. It is intuitively clear that a compact facility fits best to a hospital environment intended for particle therapy and medical diagnostics. Another criterion for selection of accelerators to be mentioned in this article is application of superconducting technology to the magnetic system design of the facility. Compact isochronous cyclotrons, which accelerate protons in the energy range 9–30 MeV, have been widely used for production of radionuclides. Energy of 230 MeV has become canonical for all proton therapy accelerators. Similar application of a carbon beam requires ion energy of 430 MeV/u. Due to application of superconducting coils the magnetic field in these machines can reach 4–5 T and even 9 T in some cases. Medical cyclotrons with an ironless or nearly ironless magnetic system that have a number of advantages over the classical accelerators are in the development stage. In this work an attempt is made to describe some conceptual and technical features of modern accelerators under consideration. The emphasis is placed on the magnetic and acceleration systems along with the beam extraction unit, which are very important from the point of view of the facility compactness and compliance with the strict medical requirements.

43 PARTICLE ACCELERATORS↗
DOE OSTI2025

Analysis Background & Noise in Stretched Wire Alignment Technique Measurements

The Stretched-Wire Alignment Technique (SWAT) is one method of magnet alignment for linear induction accelerators. The applications of SWAT have been implemented for aligning solenoid magnets on the Scorpius linear induction accelerator which will be sited at the Nevada National Security Site and the Flash X-Ray (FXR) linear induction accelerator at Lawrence Livermore National Laboratory’s Contained Firing Facility. This article describes both systematic (repeatable) and random sources of background and noise as well as practical ways to eliminate or reduce them to acceptable levels. Systematic sources include reflections from wire ends, rapid sag due to ohmic heating of the wire, magnetic materials, and shot rate. Random sources include air currents, vibration of nearby equipment, mechanical stability of test equipment, and the instruments used to measure the wire motion. Mitigations include curve fitting and adaptive noise signal cancellation, and mechanical damping. Finite Element Analysis (FEA) was used to identify and resolve a repeatable wire vibration frequency interfering with the signal resolution. Two stretched wire alignment technique set ups from Sandia National Labs and Lawrence Livermore National Lab have shown background noise sources and ways of mitigating them by either analysis methods or change of mechanical configuration. Conclusions that were drawn included the severe sensitivity of the deflection to even small external interferences of the SWAT wire such that it requires attention to detail in mechanical set up and analysis.

Linear Inductive Accelerator↗
NASA NTRS1986

The acceleration of ions and electrons by electromagnetic ion cyclotron waves

Test particle calculations of the interaction of electrons and ions with obliquely propagating electromagnetic ion cyclotron waves are presented. Such waves have frequencies below the ion cyclotron frequency. The test particle calculations show that field-aligned electrons can be accelerated parallel to the magnetic field by such waves and that ions can be accelerated perpendicular to the magnetic field to form ion conics. Ions can be accelerated much more effectively when the wave frequency is half the ion gyrofrequency. The results are applicable to acceleration of ions and electrons in the low-Beta plasma of the auroral zone.

Temerin, M.↗
DOE OSTI2019

A Potential Method for Diagnostics of Subfemtosecond Electron Beams Using Coherent Optical Transition Radiation

Coherent transition radiation (CTR) is widely used in determining the longitudinal e-beam dimensions in the accelerator technology applications. In this study, the CTR process is investigated in a real Au-target for the sake of measuring the size characteristics of the electron bunches with lengths from 100 to 300 nm, and possible applications of such targets is demonstrated.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

ICED: An Integrated CGRA Framework Enabling DFVS-Aware Acceleration

oarse-grained reconfigurable arrays (CGRAs) are a promising solution to enable energy-efficient acceleration of applications from different domains. By leveraging reconfiguration at the functional level, they can adapt to significantly different computational patterns. Existing CGRA mapping approaches extract instruction-level parallelism, exploit loop-pipelining opportunities, guarantee the data dependency, and target high throughput of a given loop. However, the recurrence data-dependency in the DFG and the mismatch between required and available computing/communication resources complicate the mapping, and might lead to significant unbalances in the utilization of the CGRA's tiles. This results in wasted power for tiles with low utilization. Applying dynamic voltage and frequency scaling (DVFS) can potentially solve this challenge and improve energy efficiency by adjusting voltage and frequency of different tiles independently. CGRAs have also been successful in accelerating data-dependent streaming applications. However, in these applications, the execution time of each kernel in the pipeline might dynamically vary depending on the characteristics of the input. This also leads to under-utilization of resources for the dynamically changing kernels that do not limit the application throughput. DVFS can also improve energy efficiency for these applications by dynamically changing the voltage and frequency levels of tiles that host non performance-constraining kernels. This paper proposes ICEDTEA -- an integrated DVFS-aware framework to map applications on CGRAs that support power islands. ICEDTEA proposes a CGRA architecture supporting DVFS islands at varying granularity (from a single tile to a group of tiles) and the related DVFS-aware compilation and mapping toolchain. ICEDTEA is the first work that introduces DVFS support for spatio-temporal CGRAs at power-island levels. The experimental evaluation shows that ICEDTEA improves average utilization by 2.3$\times$ and energy-efficiency by 1.32$\times$ over a conventional CGRA. With streaming applications, ICEDTEA improves energy efficiency by 1.12$\times$ over a state-of-the-art CGRA that introduces partial dynamic reconfiguration to adapt to variations in kernels' throughput.

Tan, Cheng↗
DOE OSTI2024

Computational study of tungsten and depleted uranium photoneutron targets for a 20 MeV electron linear accelerator

Neutron production can be realized with a high energy electron linear accelerator by using Bremsstrahlung and photoneutron converters. In this study, Monte Carlo N-Particle Code (MCNP) was used to evaluate potential photonuclear target designs for a high energy electron linear accelerator for applications such as neutron radiography and neutron resonance spectroscopy. A computational model was developed to inform a target design that would yield a high number of neutrons. It consists of a 20 MeV electron beam incident on a Bremsstrahlung target and a photonuclear target to generate neutrons. This computational model showed that a thickness of 0.75 inches for both tungsten and depleted uranium yields the most neutrons from photoneutron reactions. Saturation in the total number of generated neutrons was observed at over 0.75-inch thickness for both evaluated materials. Depleted uranium yielded approximately twice the number of neutrons overall compared to tungsten. The highest neutron surface flux for Depleted Uranium was 1.06 × 10-4 neutrons/cm2/source electron, and for Tungsten it was 5.12 × 10-5 neutrons/cm2/source electron. The optimal target design for this study’s application would consist of a 0.75 inch-thick block of depleted uranium with the length, width, and/or diameter varying dependent on application.

43 PARTICLE ACCELERATORS↗
DOE OSTI2018

Recirculating Electron Accelerators with Noncircular Electron Orbits as Radiation Sources for Applications (a Review)

State-of-the-art compact recirculating electron accelerators operating at intermediate energies (tens of MeV) are reviewed. The acceleration schemes implemented in the rhodotron, ridgetron, fantron, and cylindertron machines are discussed. Major accelerator components such as the electron guns, accelerating cavities, and bending magnets are described. The parameters of currently operating recirculating accelerators are tabulated, and applications of these accelerators in different processes of irradiation are exemplified.

43 PARTICLE ACCELERATORS↗
DOE OSTI2017

Corrigendum to “Accelerated materials evaluation for nuclear applications” [J. Nucl. Mater. 488 (2017) 46–62]

The original article addresses the opportunities and complexities of using materials test reactors with high neutron fluxes to perform accelerated studies of material aging in power reactors operating at lower neutron fluxes and with different neutron flux spectra. Radiation damage and gas production in different reactors have been compared using the code, SPECTER. This code provides a common standard from which to compare neutron damage data generated by different research groups using a variety of reactors. This Corrigendum identifies a few typographical errors. Tables 2 and 3 are included in revised form.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

Critical transverse compressive stresses of straight and bent CORC® wires with and without impregnation

CORC® wires are a promising superconductor for accelerator magnet applications. While their excellent uniaxial tensile properties have been well established, potential degradation under transverse compression remains a concern for accelerator magnets, in which transverse compression is a primary stress experienced by superconductors. To evaluate the critical transverse compressive stresses of rare-earth barium copper oxide conductors, we developed an experimental system that enables testing of samples both with and without impregnation in liquid nitrogen. Furthermore, because bending strain induced during coil winding may influence the critical compressive response of CORC® wires, the apparatus was also modified to allow testing under the bending condition. In this study wires were tested for five configurations: (1) straight, non-impregnated, (2) bent, non-impregnated, (3) straight, Stycast 2850 FT-impregnated, (4) bent, Stycast 2850 FT-impregnated, and (5) bent, paraffin wax-impregnated. The transverse pressures corresponding to 3% and 5% reductions in the critical current are reported. In conclusion, the effects of wire bending and impregnation on the critical transverse pressure are analyzed, and the implications for transverse stress levels in accelerator magnet conductors are discussed.

CORC® wire↗
DOE OSTI2022

FY2022 Status Report: Cold Spray for Canister SCC Mitigation and Repair

This progress report describes work performed during FY22 at Sandia National Laboratories (SNL) to assess the corrosion performance of cold spray coatings to enable optimization of cold spray for the purposes of mitigation and/or repair of potentially susceptible regions, corrosion, or stress corrosion cracking (SCC) in austenitic stainless steel for spent nuclear fuel (SNF) storage. Of particular concern is SCC, by which a through-wall crack could potentially form in a canister outer wall over time intervals that may be shorter than possible dry storage times. In FY21, initial corrosion explorations of cold spray coating were evaluated and in FY22, an expanded set of cold spray coatings with in-depth analysis of post-exposure accelerated testing was explored. Additionally, relevant atmospheric exposure testing was carried out and initial results are presented herein. The corrosion attack from the accelerated testing and more realistic atmospheric exposures environments were compared to identify potentially deleterious factors for corrosion as well as help to understand the applicability of accelerated testing for cold spray optimization. This initial analysis will help to enable optimization of the corrosion resistance cold spray, one of the more promising coating and repair techniques, for potential application in an SNF environment. Learnings from both are summarized, and implications and future work are presented in this report.

36 MATERIALS SCIENCE↗
DOE OSTI2022

An MLIR-based Compiler Flow for System-Level Design and Hardware Acceleration

The generation of custom hardware accelerators for applications implemented within high-level productive programming frameworks requires considerable manual effort. To automate this process, we introduce \sodaopt, a compiler tool that extends the MLIR infrastructure. \sodaopt automatically searches, outlines, tiles, and pre-optimizes relevant code regions to generate high-quality accelerators through high-level synthesis. \sodaopt can support any high-level programming framework and domain-specific language that interface with the MLIR infrastructure. By leveraging MLIR, \sodaopt solves compiler optimization problems with specialized abstractions. Backend synthesis tools connect to \sodaopt through progressive intermediate representation lowerings. \sodaopt interfaces to a design space exploration engine to identify the combination of compiler optimization passes and options that provides high-performance generated designs for different backends and targets. We demonstrate the practical applicability of the compilation flow by exploring the automatic generation of accelerators for deep neural networks operators outlined at arbitrary granularity and by combining outlining with tiling on large convolution layers. Experimental results with kernels from the PolyBench benchmark show that \sodaopt high-level optimizations improve execution delays of synthesized accelerators up to 60x. We also show that for the selected kernels, our solution outperforms the current of state-of-the art in more than 70% of the benchmarks and provides better average speedup in 55% of them.

Bohm Agostini, Nicolas↗
DOE OSTI2014

Cryogenic system for BERLinPro

In 2010 Helmholtz-Zentrum Berlin (HZB) received funding to design and build the Berlin Energy Recovery Linac Project BERLinPro. The goal of this compact Energy recovery linac (ERL) is to develop the accelerator physics and technology required to generate and accelerate a 100-mA, 1-mm mrad emittance electron beam. The BERLinPro know-how can then be transferred to various ERL-based applications. All accelerating RF cavities including the electron source are based on superconducting technology operated at 1.8 K. A Linde L700 helium liquefier is supplying 4.5 K helium. The subatmospheric pressure of 16 mbar of the helium bath of the cavities will be achieved by pumping with a set of cold compressors and warm vacuum pumps. While the L700 is already in operating, the 1.8 K system and the helium transfer system are in design phase.

43 PARTICLE ACCELERATORS↗
DOE OSTI2025

Role of Accelerated Burnup Irradiation Testing in Support of Accelerated Fuel Qualification

Accelerated fuel qualification has gained attention as a means to reduce the time needed to realize new nuclear fuel concepts and expand the operating windows of existing fuel forms. A key component of this approach is accelerated burnup irradiation testing. Although the concept of accelerated burnup has been familiar to the community for many decades, the specifics about how the increasing fission rate may be used as a qualification tool have not yet been elucidated. The present work provides a vision of how accelerated fission rate testing can enable accelerated fuel qualification. Technology readiness levels (TRLs) are reintroduced to demarcate the stages of traditional fuel qualification, and accelerated fuel qualification is presented in this context. The critical steps needed to achieve each TRL are reframed within the context of modern nuclear materials research and development, as revolutionary fuel concepts are more common than previous eras. The practical impacts of accelerated fuel qualification approaches as applied to contemporary fuel qualification efforts are illustrated. Examples are given to illustrate how accelerated burnup irradiations are being used currently and could be applied in the future to support qualification and licensure. Finally, outstanding challenges in the application of accelerated burnup methods to nuclear fuel qualification are summarized, with priority placed on understanding how fission rate impacts diffusion, microstructure evolution, and other critical mechanisms that dictate fuel performance.

Accelerated fuel qualification↗
DOE OSTI2019

Advances in global optimization of high brightness beams

High brightness electron beams play an important role in accelerator-based applications such as driving X-ray free electron laser (FEL) radiation.Here in this paper, we report on advances in global beam dynamics optimization of an accelerator design using start-to-end simulations and a new parallel multi-objective differential evolution optimization method. The global optimization results in significant improvement of the final electron beam brightness.

43 PARTICLE ACCELERATORS↗