SEARCH · Engineering Papers

Results for “Accelerator”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,567 records · Page 87

All sources NASA NTRS DOE OSTI

DOE OSTI2025

TorchBraid: High-Performance Layer-Parallel Training of Deep Neural Networks with MPI and GPU Acceleration

TorchBraid is a high-performance implementation of layer-parallel training for deep neural networks (DNNs) supporting MPI-based parallelism and GPU acceleration. Layer-parallel training has been developed to overcome the serialization inherent in forward and backward propagation of DNNs that limits utilization of computational resources in the strong scaling limit. To achieve this, TorchBraid integrates the PyTorch neural network framework with the state-of-the-art XBraid time-parallel library. Furthermore, this article presents the use and performance of TorchBraid, in addition to solutions for overcoming the algorithmic challenges inherent in combining automatic differentiation with layer-parallel. Results are presented with and without GPU acceleration for the Tiny ImageNet and MNIST image classification data sets, as well as recurrent neural networks. Overall, TorchBraid enables fast training of DNNs, both in a strong and weak scaling context. In addition to the TorchBraid software, several new advances in applying layer-parallel algorithms are detailed. Integration of layer-parallel with data-parallel algorithms is presented for the first time, showing the computational advantages of the combination. Standard deep learning techniques, like batch-normalization, are developed for layer-parallel training. Finally, a new approach combining layer-parallel with spatial coarsening in order to accelerate training for 3D image classification shows roughly a 10× speedup over serial execution.

Layer-parallel↗

DOE OSTI2020

Toward fusion plasma scenario planning for NSTX-U using machine-learning-accelerated models

One of the most promising devices for realizing power production through nuclear fusion is the tokamak. To maximize performance, it is preferable that tokamak reactors achieve advanced operating scenarios characterized by good plasma confinement, improved magnetohydrodynamic (MHD) stability, and a largely non-inductively driven plasma current. Such scenarios could enable steady-state reactor operation with high \emph{fusion gain} --- the ratio of produced fusion power to the external power provided through the plasma boundary. Precise and robust control of the evolution of the plasma boundary shape as well as the spatial distribution of the plasma current, density, temperature, and rotation will be essential to achieving and maintaining such scenarios. The complexity of the evolution of tokamak plasmas, arising due to nonlinearities and coupling between various parameters, motivates the use of model-based control algorithms that can account for the system dynamics. In this work, a learning-based accelerated model trained on data from the National Spherical Torus Experiment Upgrade (NSTX-U) is employed to develop planning and control strategies for regulating the density and temperature profile evolution around desired trajectories. The proposed model combines empirical scaling laws developed across multiple devices with neural networks trained on empirical data from NSTX-U and a database of first-principles-based computationally intensive simulations. The reduced execution time of the accelerated model will enable practical application of optimization algorithms and reinforcement learning approaches for scenario planning and control development. An initial demonstration of applying optimization approaches to the learning-based model is presented, including a strategy for mitigating the effect of leaving the finite validity range of the accelerated model. The approach shows promise for actuator planning between experiments and in real-time.

machine learning↗

DOE OSTI2024

FTTN: Feature-Targeted Testing for Numerical Properties of NVIDIA & AMD Matrix Accelerators

FTTN is a test suite to evaluate the numerical behaviors of matrix accelerators of GPUs (NVIDIA Tensor Cores and AMD Matrix Cores) in a quick and simple setting. Matrix accelerators are heavily used in today's computationally intense applications to speed up matrix multiplications. This test suite provides a comprehensive study on the numerical behaviors of these accelerators, including support for subnormals, rounding modes, extra precision bits and FMA features. Is there

Laguna Peralta, Ignacio↗

DOE OSTI2017

Impact of Poleward Moisture Transport from the North Pacific on the Acceleration of Sea Ice Loss in the Arctic since 2002

Abstract Arctic sea ice area (SIA) during late summer and early fall decreased substantially over the last four decades, and its decline accelerated beginning in the early 2000s. Statistical analyses of observations show that enhanced poleward moisture transport from the North Pacific to the Arctic Ocean contributed to the accelerated SIA decrease during the most recent period. As a consequence, specific humidity in the Arctic Pacific sector significantly increased along with an increase of downward longwave radiation beginning in 2002, which led to a significant acceleration in the decline of SIA in the Arctic Pacific sector. The resulting sea ice loss led to increased evaporation in the Arctic Ocean, resulting in a further increase of the specific humidity in mid-to-late fall, thus acting as a positive feedback to the sea ice loss. The overall set of processes is also found in a long control simulation of a coupled climate model.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2019

How Tropical Pacific Surface Cooling Contributed to Accelerated Sea Ice Melt from 2007 to 2012 as Ice is Thinned by Anthropogenic Forcing

Over the past 40 years the Arctic sea ice minimum in September has declined. The period between 2007 and 2012 showed accelerated melt contributed to the record minima of 2007 and 2012. Here, observational and model evidence shows that the changes in summer sea ice since the 2000s reflects a continuous anthropogenically forced melting masked by interdecadal variability of Arctic atmospheric circulation. This variation is partially driven by teleconnections originating from sea surface temperature (SST) changes in the east-central tropical Pacific via a Rossby wave train propagating into the Arctic (hereafter referred to as the “Pacific-Arctic teleconnection (PARC)”), which represents the leading internal mode connecting the pole to lower latitudes. This mode has contributed to accelerated warming and Arctic sea ice loss from 2007 to 2012, followed by slower declines in recent years, resulting in the appearance of a slowdown over the past 11 years. A pacemaker model simulation, in which we specify observed SST in the tropical eastern Pacific, demonstrates a physically plausible mechanism for the PARC mode. However, the model-based PARC mechanism is considerably weaker and only partially accounts for the observed acceleration of sea ice loss from 2007 to 2012. Further research is needed to identify the reasons for model limitations in reproducing the PARC mode.

Baxter, Ian↗

DOE OSTI2017

A GPU-accelerated continuous and discontinuous Galerkin non-hydrostatic atmospheric model

In the article, we present a Graphics Processing Unit (GPU)-accelerated nodal discontinuous Galerkin method for the solution of the three-dimensional Euler equations that govern the motion and thermodynamic state of the atmosphere. Acceleration of the dynamical core of atmospheric models plays an important practical role in not only getting daily forecasts faster, but also in obtaining more accurate (high resolution) results within a given simulation time limit. We use algorithms suitable for the single instruction multiple thread architecture of GPUs to accelerate our model by two orders of magnitude relative to one core of a CPU. Tests on one node of the Titan supercomputer show a speedup of up to 15 times using the K20X GPU as compared to that on the 16-core AMD Opteron CPU. The scalability of the multi-GPU implementation is tested using 16,384 GPUs, which resulted in a weak scaling efficiency of about 90%. Finally, the accuracy and performance of our GPU implementation is verified using several benchmark problems representative of different scales of atmospheric dynamics.

Computer Science↗

DOE OSTI2023

HipBone: A performance-portable graphics processing unit-accelerated C++ version of the NekBone benchmark

We present hipBone, an open-source performance-portable proxy application for the Nek5000 (and NekRS) computational fluid dynamics applications. HipBone is a fully GPU-accelerated C++ implementation of the original NekBone CPU proxy application with several novel algorithmic and implementation improvements which optimize its performance on modern fine-grain parallel GPU accelerators. Our optimizations include a conversion to store the degrees of freedom of the problem in assembled form in order to reduce the amount of data moved during the main iteration and a portable implementation of the main Poisson operator kernel. We demonstrate near-roofline performance of the operator kernel on three different modern GPU accelerators from two different vendors. We present a novel algorithm for splitting the application of the Poisson operator on GPUs which aggressively hides MPI communication required for both halo exchange and assembly. Our implementation of nearest-neighbor MPI communication then leverages several different routing algorithms and GPU-Direct RDMA capabilities, when available, which improves scalability of the benchmark. We demonstrate the performance of hipBone on three different clusters housed at Oak Ridge National Laboratory, namely, the Summit supercomputer and the Frontier early-access clusters, Spock and Crusher. Our tests demonstrate both portability across different clusters and very good scaling efficiency, especially on large problems.

Computer Science↗

DOE OSTI2024

An FPGA-based hardware accelerator supporting sensitive sequence homology filtering with profile hidden Markov models

Abstract Background Sequence alignment lies at the heart of genome sequence annotation. While the BLAST suite of alignment tools has long held an important role in alignment-based sequence database search, greater sensitivity is achieved through the use of profile hidden Markov models (pHMMs). Here, we describe an FPGA hardware accelerator, called HAVAC, that targets a key bottleneck step (SSV) in the analysis pipeline of the popular pHMM alignment tool, HMMER. Results The HAVAC kernel calculates the SSV matrix at 1739 GCUPS on a $$\sim$$ ∼ $3000 Xilinx Alveo U50 FPGA accelerator card, $$\sim$$ ∼ 227× faster than the optimized SSV implementation in nhmmer . Accounting for PCI-e data transfer data processing, HAVAC is 65× faster than nhmmer’s SSV with one thread and 35× faster than nhmmer with four threads, and uses $$\sim$$ ∼ 31% the energy of a traditional high end Intel CPU. Conclusions HAVAC demonstrates the potential offered by FPGA hardware accelerators to produce dramatic speed gains in sequence annotation and related bioinformatics applications. Because these computations are performed on a co-processor, the host CPU remains free to simultaneously compute other aspects of the analysis pipeline.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2018

Design of a tapered slot waveguide dielectric laser accelerator for sub-relativistic electrons

We propose a dielectric laser accelerator design based on a tapered slot waveguide structure for sub-relativistic electron acceleration. This tapering scheme allows for straightforward tuning of the phase velocity of the accelerating field along the propagation direction, which is necessary for maintaining synchronization with electrons as their velocities increase. Furthermore, the non-resonant nature of this design allows for better tolerance to experimental errors. We also introduce a method to design this continuously tapered structure based on the eikonal approximation, and give a working example based on realistic experimental parameters.

47 OTHER INSTRUMENTATION↗

DOE OSTI2017

WH2 and proline-rich domains of WASP-family proteins collaborate to accelerate actin filament elongation

WASP-family proteins are known to promote assembly of branched actin networks by stimulating the filament-nucleating activity of the Arp2/3 complex. Here, we show that WASP-family proteins also function as polymerases that accelerate elongation of uncapped actin filaments. When clustered on a surface, WASP-family proteins can drive branched actin networks to grow much faster than they could by direct incorporation of soluble monomers. This polymerase activity arises from the coordinated action of two regulatory sequences: (i) a WASP homology 2 (WH2) domain that binds actin, and (ii) a proline-rich sequence that binds profilin–actin complexes. In the absence of profilin, WH2 domains are sufficient to accelerate filament elongation, but in the presence of profilin, proline-rich sequences are required to support polymerase activity by (i) bringing polymerization-competent actin monomers in proximity to growing filament ends, and (ii) promoting shuttling of actin monomers from profilin–actin complexes onto nearby WH2 domains. Unoccupied WH2 domains transiently associate with free filament ends, preventing their growth and dynamically tethering the branched actin network to the WASP-family proteins that create it. Collaboration between WH2 and proline-rich sequences thus strikes a balance between filament growth and tethering. Our work expands the number of critical roles that WASP-family proteins play in the assembly of branched actin networks to at least three: (i) promoting dendritic nucleation; (ii) linking actin networks to membranes; and (iii) accelerating filament elongation.

Biochemistry & Molecular Biology↗

DOE OSTI2018

Speculation and replication in temperature accelerated dynamics

Accelerated Molecular Dynamics (AMD) is a class of MD-based algorithms for the long-time scale simulation of atomistic systems that are characterized by rare-event transitions. Temperature-Accelerated Dynamics (TAD), a traditional AMD approach, hastens state-to-state transitions by performing MD at an elevated temperature. Recently, Speculatively-Parallel TAD (SpecTAD) was introduced, allowing the TAD procedure to exploit parallel computing systems by concurrently executing in a dynamically generated list of speculative future states. Although speculation can be very powerful, it is not always the most efficient use of parallel resources. In this paper, we compare the performance of speculative parallelism with a replica-based technique, similar to the Parallel Replica Dynamics method. A hybrid SpecTAD approach is also presented, in which each speculation process is further accelerated by a local set of replicas. Finally and overall, this work motivates the use of hybrid parallelism whenever possible, as some combination of speculation and replication is typically most efficient.

36 MATERIALS SCIENCE↗

DOE OSTI2022

Is Mini Beam Ready for Human Trials? Results of Randomized Study of Treating De-Novo Brain Tumors in Canines Using Linear Accelerator Generated Mini Beams

The main challenge in treating malignant brain neoplasms lies in eradicating the tumor while minimizing treatmentrelated damage. Conventional radiation treatments are associated with considerable side effects. Synchrotron generated micro-beam radiation (SMBRT) has shown to preserve brain architecture while killing tumor cells, however physical characteristics and limited facility access restrict its use. We have created a new clinical device which produces mini beams on a linear accelerator, to provide a new type of treatment called mini-beam radiation therapy (MBRT). The objective of this study is to compare the treatment outcomes of linear accelerator based MBRT versus standard radiation treatment (SRT), to evaluate the tumor response and the treatment-related changes in the normal brain with respect to each treatment type. Pet dogs with de-novo brain tumors were accrued for treatment. Dogs were randomized between standard fractionated stereotactic (9 Gy in 3 fractions) radiation treatment vs. a single fraction of MBRT (26 Gy mean dose). Dogs were monitored after treatment for clinical assessment and imaging. When the dogs were euthanized, a veterinary pathologist assessed the radiation changes and tumor response. We accrued 16 dogs, 8 dogs in each treatment arm. In the MBRT arm, 71% dogs achieved complete pathological remission. The radiation-related changes were all confined to the target region. Structural damage was not observed in the beam path outside of the target region. In contrast, none of the dogs in control group achieved remission and the treatment related damage was more extensive. Therapeutic superiority was observed with MBRT, including both tumor control and the normal structural preservation. The MBRT findings are suggestive of an immune related mechanism which is absent in standard treatment. These findings together with the widespread availability of clinical linear accelerators make MBRT a promising research topic to explore further treatment and clinical trial opportunities.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2023

TEACHING AN OLD ACCELERATOR NEW TRICKS

The Argonne Tandem Linac Accelerator System (ATLAS) has been a National User Facility since 1985. In that time, many of the systems that help operators retrieve, modify, and store beamline parameters have not kept pace with the advancement of technology. Development of a new method of storing and retrieving beamline parameters resulted in the testing and installation of a time-series database as a potential replacement for the traditional relational database. InfluxDB was selected due to its self-hosted Open-Source version availability as well as the simplicity of installation and setup. A program was written to periodically gather all accelerator parameters in the control system and store them in the time-series database. This resulted in over 13,000 distinct data points, captured at 5-minute intervals. A second test captured 35 channels on a 1-minute cadence. Graphing of the captured data is being done on Grafana, an Open-Source version is available that co-exists well with InfluxDB as the back-end. Grafana made visualizing the data simple and flexible. The testing has allowed for the use of modern graphing tools to generate new insights into operating the accelerator, as well as opened the door to building large data sets suitable for Artificial Intelligence and Machine Learning applications.

Novak, D.↗

DOE OSTI2022

Need for portable accelerators in cultural heritage

Ion Beam Accelerators (IBA) centres have provided researchers with powerful techniques to analyse objects of cultural significance in a non-destructive and non-invasive manner. However, in some cases it is not feasible to remove an object from the field or museum and transport it to the laboratory. In this contributed talk, we present as a manner of a short review, examples of the benefits provided from these techniques in the study of material culture and discuss the initial steps to consider when investigating the feasibility of a compact accelerator that can be taken to sites of cultural significance for PIXE analysis. In particular, we consider the application of a compact, robust 2 MeV proton accelerator that can be taken into the field to perform PIXE measurements on rock art. We detail the main challenges and considerations for such a device.

Charles, T. K.↗

DOE OSTI2019

RF Results of Nb Coated SRF Accelerator Cavities via HiPIMS

Bulk Niobium (Nb) SRF (superconducting radio frequency) cavities are currently the preferred method for acceleration of charged particles at accelerator facilities around the world. However, bulk Nb cavities have poor thermal conductance and impose material and design restrictions on other components of a particle accelerator. Since the SRF phenomena occurs within a shallow depth of 40 nm (for Nb), a proposed solution to this problem has been to deposit a superconducting Nb thin film on the interior of a cavity made of a suitable alternative material such as copper or aluminum. While this approach has been attempted in the past using DC magnetron sputtering (DCMS), such cavities have never performed at the bulk Nb level. However, new energetic condensation techniques for film deposition offer the opportunity to create suitably thick Nb films with improved density, microstructure and adhesion compared to traditional DCMS. One such technique that has been developed somewhat recently is ?High Power Impulse Magnetron Sputtering? (HiPIMS). Here we report early results from various thin film coatings carried out on 1.3 GHz Cu cavities, a 1.5 GHz Nb cavity, and small Cu coupon samples coated at Jefferson Lab using HiPIMS.

Burton, Matthew Creed↗

DOE OSTI2025

Beam optics design of a prototype 20 kW conduction-cooled SRF accelerator for medical sterilization

Superconducting technology has significantly advanced the capabilities of particle accelerators, facilitating higher beam-power operations for fundamental research at a comparatively lower cost. However, the conventional implementation of superconducting technology introduces complexities in the form of cryogenic plants, cryogenic distribution systems and substantial construction and operational cost. In response to these challenges, recent research efforts at Fermilab have been dedicated to the development of a cryogen-free, conduction-cooled Nb3Sn-based superconducting technology. This paper outlines the beam optics design of a 20-kW conduction-cooled compact superconducting accelerator for medical sterilization. The paper reviews both the physics and practical constraints associated with high beam-power operation within the context of industrial applications. The focus is on providing insights into the potential of this innovative technology to overcome existing challenges and pave the way for more accessible and efficient industrial particle accelerators.

Saini, A. [Fermilab]↗

DOE OSTI2019

RECENT DEVELOPMENTS OF Nb3Sn AT JEFFERSON LAB FOR SRF ACCELERATOR APPLICATION

The desire to reduce the construction and operating costs of future SRF accelerators motivates the search for alternative, higher-performing materials. Nb3Sn (Tc ~ 18.3 K and Hsh ~ 425 mT) is the front runner. However, tests of early Nb3Sn-coated cavities encountered strong Q-slopes limiting the performance. Learnings from studies of coated materials related to cavity performance prompted significant changes to the coating process. It is now possible to routinely produce slope-free cavities having Q0 > 2×1010 at 4 K and > 4×1010 up to the accelerating gradient in excess of 15 MV/m at its best. Obtaining similar results in five-cell cavities is a current goal to test them under an accelerator environment. This contribution discusses recent developments at Jefferson Lab.

Pudasaini, Uttar↗

DOE OSTI2017

ALON® Components With Tunable Dielectric Properties for High Power Accelerator Applications

There are challenges in linear particle accelerators associated with the need to suppress “higher order modes” (HOMs). HOMs are detrimental to accelerator operation as they are a source of beam instability. The absorption/suppression of HOMs and dissipation of the energy of higher order modes is vital to the function of these accelerators. Surmet has identified ALON® Optical Ceramic (Aluminum Oxynitride), a hard, durable ceramic that is fabricated through conventional powder processing techniques, as a potential material for HOM absorber. In this Phase I program, Surmet has produced new ALON-composite HOM absorber materials that function at both ambient and cryogenic temperatures. The composite materials were developed and evaluated in collaboration with Thomas Jefferson National Labs. Success in this Phase I and the potential Phase II will demonstrate the utility of ALON composite components for RF absorbing applications and lay the groundwork for commercialization of such products, with applications in basic science, medical and digital electronics industries.

36 MATERIALS SCIENCE↗