Search NASA⌕ Search

SEARCH · Search NASA

Results for “Accelerators, 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 1,333 records · Page 74

Inference-Optimized AI and High Performance Computing for Gravitational Wave Detection at Scale

We introduce an ensemble of artificial intelligence models for gravitational wave detection that we trained in the Summit supercomputer using 32 nodes, equivalent to 192 NVIDIA V100 GPUs, within 2 h. Once fully trained, we optimized these models for accelerated inference using NVIDIA TensorRT. We deployed our inference-optimized AI ensemble in the ThetaGPU supercomputer at Argonne Leadership Computer Facility to conduct distributed inference. Using the entire ThetaGPU supercomputer, consisting of 20 nodes each of which has 8 NVIDIA A100 Tensor Core GPUs and 2 AMD Rome CPUs, our NVIDIA TensorRT-optimized AI ensemble processed an entire month of advanced LIGO data (including Hanford and Livingston data streams) within 50 s. Our inference-optimized AI ensemble retains the same sensitivity of traditional AI models, namely, it identifies all known binary black hole mergers previously identified in this advanced LIGO dataset and reports no misclassifications, while also providing a 3X inference speedup compared to traditional artificial intelligence models. We used time slides to quantify the performance of our AI ensemble to process up to 5 years worth of advanced LIGO data. In this synthetically enhanced dataset, our AI ensemble reports an average of one misclassification for every month of searched advanced LIGO data. We also present the receiver operating characteristic curve of our AI ensemble using this 5 year long advanced LIGO dataset. This approach provides the required tools to conduct accelerated, AI-driven gravitational wave detection at scale.

97 MATHEMATICS AND COMPUTING↗

HEP Science Network Requirements--Final Report

The Energy Sciences Network (ESnet) is the primary provider of network connectivity for the US Department of Energy Office of Science, the single largest supporter of basic research in the physical sciences in the United States. In support of the Office of Science programs, ESnet regularly updates and refreshes its understanding of the networking requirements of the instruments, facilities, scientists, and science programs that it serves. This focus has helped ESnet to be a highly successful enabler of scientific discovery for over 20 years. In August 2009 ESnet and the Office of High Energy Physics (HEP), of the DOE Office of Science, organized a workshop to characterize the networking requirements of the programs funded by HEP. The International HEP community has been a leader in data intensive science from the beginning. HEP data sets have historically been the largest of all scientific data sets, and the communty of interest the most distributed. The HEP community was also the first to embrace Grid technologies. The requirements identified at the workshop are summarized below, and described in more detail in the case studies and the Findings section: (1) There will be more LHC Tier-3 sites than orginally thought, and likely more Tier-2 to Tier-2 traffic than was envisioned. It it not yet known what the impact of this will be on ESnet, but we will need to keep an eye on this traffic. (2) The LHC Tier-1 sites (BNL and FNAL) predict the need for 40-50 Gbps of data movement capacity in 2-5 years, and 100-200 Gbps in 5-10 years for HEP program related traffic. Other key HEP sites include LHC Tier-2 and Tier-3 sites, many of which are located at universities. To support the LHC, ESnet must continue its collaborations with university and international networks. (3) While in all cases the deployed 'raw' network bandwidth must exceed the user requirements in order to meet the data transfer and reliability requirements, network engineering for trans-Atlantic connectivity is more complex than network engineering for intra-US connectivity. This is because transoceanic circuits have lower reliability and longer repair times when compared with land-based circuits. Therefore, trans-Atlantic connectivity requires greater deployed bandwidth and diversity to ensure reliability and service continuity of the user-level required data transfer rates. (4) Trans-Atlantic traffic load and patterns must be monitored, and projections adjusted if necessary. There is currently a shutdown planned for the LHC in 2012 that may affect projections of trans-Atlantic bandwidth requirements. (5) There is a significant need for network tuning and troubleshooting during the establishment of new LHC Tier-2 and Tier-3 facilities. ESnet will work with the HEP community to help new sites effectively use the network. (6) SLAC is building the CCD camera for the LSST. This project will require significant bandwidth (up to 30Gbps) to NCSA over the next few years. (7) The accelerator modeling program at SLAC could require the movement of 1PB simulation data sets from the Leadership Computing Facilities at Argonne and Oak Ridge to SLAC. The data sets would need to be moved overnight, and moving 1PB in eight hours requires more than 300Gbps of throughput. This requirement is dependent on the deployment of analysis capabilities at SLAC, and is about five years away. (8) It is difficult to achieve high data transfer throughput to sites in China. Projects that need to transfer data in or out of China are encouraged to deploy test and measurement infrastructure (e.g. perfSONAR) and allow time for performance tuning.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Graph Neural Network-based Tracking as a Service

Recent studies have shown promising results for track finding in dense environments using Graph Neural Network (GNN)-based algorithms. However, GNN-based track finding is computationally slow on CPUs, necessitating the use of coprocessors to accelerate the inference time. Additionally, the large input graph size demands a large device memory for efficient computation, a requirement not met by all computing facilities used for particle physics experiments, particularly those lacking advanced GPUs. Furthermore, deploying the GNN-based track-finding algorithm in a production environment requires the installation of all dependent software packages, exclusively utilized by this algorithm. These computing challenges must be addressed for the successful implementation of GNN-based track-finding algorithm into production settings. In response, we introduce a ``GNN-based tracking as a service'' approach, incorporating a custom backend within the NVIDIA Triton inference server to facilitate GNN-based tracking. This paper presents the performance of this approach using the Perlmutter supercomputer at NERSC.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Contact detection acceleration in pebble flow simulation for pebble bed reactor systems

Pebble flow simulation plays an important role in the steady state and transient analysis of thermal-hydraulics and neutronics for Pebble Bed Reactors (PBR). The Discrete Element Method (DEM) and the modified Molecular Dynamics (MD) method are widely used to simulate the pebble motion to obtain the distribution of pebble concentration, velocity, and maximum contact stress. Although DEM and MD present high accuracy in the pebble flow simulation, they are quite computationally expensive due to the large quantity of pebbles to be simulated in a typical PBR and the ubiquitous contacts and collisions between neighboring pebbles that need to be detected frequently in the simulation, which greatly restricted their applicability for large scale PBR designs such as PBMR400. Since the contact detection accounts for more than 60% of the overall CPU time in the pebble flow simulation, the acceleration of the contact detection can greatly enhance the overall efficiency. In the present work, based on the design features of PBRs, two contact detection algorithms, the basic cell search algorithm and the bounding box search algorithm are investigated and applied to pebble contact detection. The influence from the PBR system size, core geometry and the searching cell size on the contact detection efficiency is presented. Our results suggest that for present PBR applications, the bounding box algorithm is less sensitive to the aforementioned effects and has superior performance in pebble contact detection compared with basic cell search algorithm. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Prediction of Adsorption Energies for Chemical Species on Metal Catalyst Surfaces Using Machine Learning

Computational catalyst screening has the potential to significantly accelerate heterogeneous catalyst discovery. Typically, this involves developing microkinetic reactor models that are based on parameters obtained from density functional theory and transition-state theory. To reduce the large computational cost involved in computing various adsorption and transition-state energies of all possible surface states on a large number of catalyst models, linear scaling relations for surface intermediates and transition states have been developed that only depend on a few, typically one or two descriptors, such as the carbon atom adsorption energy. As a result, only the descriptor values have to be computed for various active site models to generate volcano curves in activity or selectivity. Unfortunately, for more complex chemistries the predictability of linear scaling relations is unknown. Also, the selection of descriptors is essentially a trial and error process. Here, using a database of adsorption energies of the surface species involved in the decarboxylation and decarbonylation of propionic acid over eight monometalic transition-metal catalyst surfaces (Ni, Pt, Pd, Ru, Rh, Re, Cu, Ag), we tested if nonlinear machine learning (ML) models can outperform the linear scaling relations in prediction accuracy when predicting the adsorption energy for various species on a metal surface based on data from the rest of the metal surfaces. We found linear scaling relations to hold well for predictions across metals with a mean-absolute error of 0.12 eV, and ML methods being unable to outperform linear scaling relations when the training dataset contains a complete set of energies for all of the species on various metal surfaces. Only when the training dataset is incomplete, namely, contains a random subset of species’ energies for each metal, a currently unlikely scenario for catalyst screening, do kernel-based ML models significantly outperform linear scaling relations. We also found that simple coordinate-free species descriptors, such as bond counts, achieve as good results as sophisticated coordinate-based descriptors. Finally, we propose an approach for automatic discovery of appropriate metal descriptors using principal component analysis.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Modeling High-Energy Gamma-Rays from the Fermi Bubbles - Oral Presentation [Slides]

In 2010, the Fermi Bubbles were discovered at the galactic center of the Milky Way. These giant gamma-ray structures, extending 50 degrees in galactic latitude and 20-30 degrees in galactic longitude, were not predicted. We wish to develop a model for the gamma-ray emission of the Fermi Bubbles. To do so, we assume that second order Fermi acceleration is responsible for the high-energy emission of the bubbles. Second order Fermi acceleration requires charged particles and irregular magnetic fields—both of which are present in the disk of the Milky Way galaxy. I use the assumption of second order Fermi acceleration in the transport equation, which describes the diffusion of particles. By solving the steady-state case of the transport equation, I compute the proton spectrum due to Fermi second order acceleration and compare this analytical solution to a numerical solution provided by Dr. P. Mertsch. Analytical solutions to the transport equation are taken from Becker, Le, & Dermer and are used to further test the numerical solution. I find that the numerical solution converges to the analytical solution in all cases. Thus, we know the numerical solution accurately calculates the proton spectrum. The gamma-ray spectrum follows the proton spectrum, and will be computed in the future.

79 ASTRONOMY AND ASTROPHYSICS↗

White Paper on Nuclear Astrophysics and Low Energy Nuclear Physics - Part 1. Nuclear Astrophysics

This white paper informs the nuclear astrophysics community and funding agencies about the scientific directions and priorities of the field and provides input from this community for the 2015 Nuclear Science Long Range Plan. It summarizes the outcome of the nuclear astrophysics town meeting that was held on August 21 - 23, 2014 in College Station at the campus of Texas A&M University in preparation of the NSAC Nuclear Science Long Range Plan. It also reflects the outcome of an earlier town meeting of the nuclear astrophysics community organized by the Joint Institute for Nuclear Astrophysics (JINA) on October 9 - 10, 2012 Detroit, Michigan, with the purpose of developing a vision for nuclear astrophysics in light of the recent NRC decadal surveys in nuclear physics (NP2010) and astronomy (ASTRO2010). The white paper is furthermore informed by the town meeting of the Association of Research at University Nuclear Accelerators (ARUNA) that took place at the University of Notre Dame on June 12 - 13, 2014. In summary we find that nuclear astrophysics is a modern and vibrant field addressing fundamental science questions at the intersection of nuclear physics and astrophysics. These questions relate to the origin of the elements, the nuclear engines that drive life and death of stars, and the properties of dense matter. A broad range of nuclear accelerator facilities, astronomical observatories, theory efforts, and computational capabilities are needed. With the developments outlined in this white paper, answers to long-standing key questions are well within reach in the coming decade.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

CFD Simulation of Aerobic Gas Fermentation to Enable Commercial Conversion of CO 2 into Aquaculture and Animal Feed: Cooperative Research and Development (Final Report)

NovoNutrients’ fermentation technology uses energy from hydrogen to transform industrial CO2 emissions into premium animal feed ingredients and other valuable products. A single NovoNutrients’ commercial manufacturing plant will capture and convert over 200,000 tons/yr of CO2 into over 100,000 tons/yr of high-protein feed. Key to the rapid and widespread deployment of the technology is maximization of its productivity and energy efficiency. Robust, physically based computational models of the technology will significantly increase productivity and efficiency, accelerating NovoNutrients’ technology to manufacturing scale. NREL has unique capabilities for creating and running such computational models. NREL's existing aerobic bioreaction computational fluid dynamics (CFD) models will be adapted to NovoNutrients’ gas fermentation (CO2, H2, O2) technology. The multiphysics CFD simulations require thousands of high-performance computing (HPC) node hours to simulate the complex geometries and contents of NovoNutrients’ industrial bioreactors. The experimentally validated CFD models were used to identify optimally efficient and productive bioreactor designs and operating conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

areaDetector: Software for 2-D Detectors in EPICS

areaDetector is a new EPICS module designed to support 2-D detectors. It is modular C++ code that greatly simplifies the task of writing support for a new detector. It also supports plugins, which receive detector data from the driver and process it in some way. Existing plugins perform Region-Of-Interest extraction and analysis, file saving (in netCDF, HDF, TIFF and JPEG formats), color conversion, and export to EPICS records for image display in clients like ImageJ and IDL. Drivers have now been written for many of the detectors commonly used at synchrotron beamlines, including CCDs, pixel array and amorphous silicon detectors, and online image plates.

99 GENERAL AND MISCELLANEOUS↗

XACC - eXtreme-scale Accelerator Programming Framework

Hybrid programming models for beyond-CMOS technologies will prove critical for integrating new computing technologies alongside our existing infrastructure. Unfortunately the software infrastructure required to enable this is lacking or not available. XACC is a programming framework for extreme-scale, post-exascale accelerator architectures that integrates alongside existing conventional applications. It is a pluggable framework for programming languages developed for next-gen computing hardware architectures like quantum and neuromorphic computing. It lets computational scientists efficiently off-load classically intractable work to attached accelerators through user-friendly Kernel definitions. XACC makes post-exascale hybrid programming approachable for domain computational scientists.

McCaskey, AlexanderJ.↗

XACC - eXtreme-scale Accelerator Programming Framework

Hybrid programming models for beyond-CMOS technologies will prove critical for integrating new computing technologies alongside our existing infrastructure. Unfortunately the software infrastructure required to enable this is lacking or not available. XACC is a programming framework for extreme-scale, post-exascale accelerator architectures that integrates alongside existing conventional applications. It is a pluggable framework for programming languages developed for next-gen computing hardware architectures like quantum and neuromorphic computing. It lets computational scientists efficiently off-load classically intractable work to attached accelerators through user-friendly Kernel definitions. XACC makes post-exascale hybrid programming approachable for domain computational scientists.

McCaskey, Alexander J.↗

Simple formula for surface roughness wakes

For energy recovery linacs (ERLs) the energy spread induced by wake- fields grows throughout acceleration and decelleration. Too much fractional momentum spread at low energy will complicate beam transport. This note uses some analytical results and a simple model of surface roughness correlation functions to obtain a wakefield formula that is very fast and easy to compute.

43 PARTICLE ACCELERATORS↗

Long-range corrected density functional theory with accelerated Hartree-Fock exchange integration using a two-Gaussian operator [LC-ωPBE(2Gau)]

Since the advent of hybrid functional in 1993, it has become a main quantum chemical tool for the calculation of energies and properties of molecular systems. Following the introduction of long-range corrected hybrid scheme for density functional theory a decade later, the applicability of the hybrid functional has been further amplified due to the resulting increased performance on orbital energy, excitation energy, non-linear optical property, barrier height, and so on. Nevertheless, the high cost associated with the evaluation of Hartree-Fock (HF) exchange integrals remains a bottleneck for the broader and more active applications of hybrid functionals to large molecular and periodic systems. Here, we propose a very simple yet efficient method for the computation of long-range corrected hybrid scheme. It uses a modified two-Gaussian attenuating operator instead of the error function for the long-range HF exchange integral. As a result, the two-Gaussian HF operator, which mimics the shape of the error function operator, reduces computational time dramatically (e.g., about 14 times acceleration in C diamond calculation using periodic boundary condition) and enables lower scaling with system size, while maintaining the improved features of the long-range corrected density functional theory.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A hyper-spherical adaptive sparse-grid method for high-dimensional discontinuity detection

This work proposes and analyzes a hyper-spherical adaptive hierarchical sparse-grid method for detecting jump discontinuities of functions in high-dimensional spaces is proposed. The method is motivated by the theoretical and computational inefficiencies of well-known adaptive sparse-grid methods for discontinuity detection. Our novel approach constructs a function representation of the discontinuity hyper-surface of an N-dimensional dis- continuous quantity of interest, by virtue of a hyper-spherical transformation. Then, a sparse-grid approximation of the transformed function is built in the hyper-spherical coordinate system, whose value at each point is estimated by solving a one-dimensional discontinuity detection problem. Due to the smoothness of the hyper-surface, the new technique can identify jump discontinuities with significantly reduced computational cost, compared to existing methods. Moreover, hierarchical acceleration techniques are also incorporated to further reduce the overall complexity. Rigorous error estimates and complexity analyses of the new method are provided as are several numerical examples that illustrate the effectiveness of the approach.

97 MATHEMATICS AND COMPUTING↗

Accelerated Monte Carlo Fission Source Convergence with Fission Matrix and Kernel Density Estimators

Accurate solution of the k-eigenvalue form of the Boltzmann radiation transport equation is required in many application areas, including nuclear reactor analysis and criticality safety. Monte Carlo calculations are free of the discretization errors inherent to deterministic methods and therefore offer the most accurate solutions. However, converging the fission source can be difficult for many problems, resulting in either a substantial computational overhead or potential inaccuracies that negate many of the benefits of using Monte Carlo methods. Difficulty in converging the fission source occurs for two distinct reasons. The first is the existence of a high dominance ratio in the problem (i.e. the largest and second largest eigenvalues have approximately the same magnitude). The second difficulty arises due to undersampling of the fission source because an insufficient number of particle histories are simulated. The first difficulty is shared by deterministic solvers when using power iteration, while the second difficulty is unique to Monte Carlo solvers. In practice, both difficulties may be present in the same problem, compounding the difficulty in achieving a converged distribution. In this paper, we consider a new approach to accelerate the convergence of the fission source based on combining the use of fission matrix methods with kernel density estimators (KDEs). It has been observed that deterministic fission matrix methods show the potential to accelerate the convergence of the fission source in Monte Carlo criticality calculations but may introduce significant additional noise for difficult problems. In this paper, we have demonstrated the use of a KDE approach in combination with the fission matrix acceleration. Although the addition of the KDE initially slows the convergence of the fission matrix acceleration, it decreases the magnitude of the noise while still offering significantly improved convergence relative to analog Monte Carlo. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

HYDJET++ model of relativistic heavy-ion collisions

The HYDJET++ event generator intended for simulating multiparticle hadron production in relativistic heavy-ion collisions over a broad energy range ( sqrt s {approx} 10 - 10 000 GeV per nucleon pair) is described in the present article. Within HYDJET++, the final state of a nuclear reaction is a superposition of two independent components: a soft, hydrodynamic, part (processes involving the production of low-transverse-momentum hadrons) and a hard, multiparton, state (processes involving the production of high-transverse-momentum hadrons via the fragmentation of quark-gluon jets). Some results of a simulation of nucleus-nucleus interactions on the basis of the HYDJET++ event generator at the energies of the Relativistic Heavy-Ion Collider (RHIC) and the Large Hadron Collider (LHC) are discussed.

43 PARTICLE ACCELERATORS↗