Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallelization efficiency”

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 1,765 records · Page 98

Parallel Particle Advection Bake-Off for Scientific Visualization Workloads

There are multiple algorithms for parallelizing particle advection for scientific visualization workloads. While many previous studies have contributed to the understanding of individual algorithms, our study aims to provide a holistic understanding of how algorithms perform relative to each other on various workloads. To accomplish this, we consider four popular parallelization algorithms and run a “bake-off” study (i.e., an empirical study) to identify the best matches for each. The study includes 216 tests, going to a concurrency of up to 8192 cores and considering data sets as large as 34 billion cells with 300 million particles. Overall, our study informs three important research questions: (1) which parallelization algorithms perform best for a given workload?, (2) why?, and (3) what are the unsolved problems in parallel particle advection? In terms of findings, we find that the seeding box is the most important factor in choosing the best algorithm, and also that there is a significant opportunity for improvement in execution time, scalability, and efficiency.

Pugmire, Dave↗

Investigation of Portable Event-Based Monte Carlo Transport Using the NVIDIA Thrust Library

Power consumption considerations are driving future high performance computing platforms toward many-core computing architectures. Los Alamos National Laboratory's Trinity machine, available in 2016, will use both Intel Xeon Haswell processors and Intel Xeon Phi Knights Landing many integrated core (MIC) architecture coprocessors. Lawrence Livermore National Laboratory's Sierra machine, available in 2018, will use an IBM PowerPC architecture along with Nvidia graphics processing unit (GPU) architecture accelerators. These different advanced architectures make the computing landscape in upcoming years complex. Traditional approaches to Monte Carlo transport do not work efficiently on these new computing platforms. MIC architectures require vectorization to operate efficiently, and vectorization is difficult to achieve in Monte Carlo transport. GPU architectures require additional code to explicitly use the hardware, requiring significant code changes or hardware specific branches in the source code. A significant challenge for Monte Carlo transport projects is to simultaneously support within a single source code base efficient simulations for both the current generation of architectures and the different advanced computing architectures. In order to address these challenges, two important changes are typically required: a new algorithmic approach for solving Monte Carlo transport, and explicit use of hardware specific software. In this paper, we describe initial research investigations of an event-based Monte Carlo transport algorithm implemented using the Nvidia Thrust library on a GPU for a Monte Carlo test code. The event-based algorithm targets many-core architectures by increasing SIMD (single instruction multiple data) parallelism, while Thrust potentially provides portable performance by allowing one source code base to compile code targeted for both CPUs and GPUs. We described preliminary investigations of portable event-based Monte Carlo algorithms implemented using the Nvidia Thrust library in a research Monte Carlo test code. We found that an explicit CUDA implementation of an event-based Monte Carlo algorithm performed significantly more efficiently than a Thrust implementation on GPU platforms, most likely as a result of additional flexibility in access to different memory spaces on the GPU. Additionally, we showed that on GPU platforms and at large enough problem sizes the event-based implementations perform more efficiently than the serial history-based implementation running on the host CPU. While investigating this problem, we also discovered that the performance of the event-based algorithm is affected by what tallies are being used. A zonal scalar flux tally requires atomic operations that significantly impacted the performance of the code, in some cases producing slowdowns instead of speedups. We decided to remove the tally in order to focus on the effectiveness of the event-based algorithm. Future work will be required to research more effective ways of handling such tallies. Additionally, we would like to consider new ways for optimizing both the Thrust and CUDA versions, in order to see how much performance we have yet to achieve. The potential trade-off between portability and performance was demonstrated in this investigation. Thrust provides both CPU and GPU versions of the code in one code base, but it does so at a cost. We discovered approximately a factor of 2-5 performance difference between Thrust and CUDA on each GPU platform we tested. Thrust provides a tool to access the GPUs with less effort and specialization, but it does so by giving up fine grained control where extra performance can be found for this application. Future work will be required to determine whether this performance differential between Thrust and CUDA can be reduced for event-based Monte Carlo transport. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Directional multimode coupler for planar magnonics: Side-coupled magnetic stripes

We experimentally demonstrate spin waves coupling in two laterally adjacent magnetic stripes. By the means of Brillouin light scattering spectroscopy, we show that the coupling efficiency depends both on the magnonic waveguides' geometry and the characteristics of spin-wave modes. In particular, the lateral confinement of coupled yttrium-iron-garnet stripes enables the possibility of control over the spin-wave propagation characteristics. Numerical simulations (in time domain and frequency domain) reveal the nature of intermodal coupling between two magnonic stripes. The proposed topology of multimode magnonic coupler can be utilized as a building block for fabrication of integrated parallel functional and logic devices such as the frequency selective directional coupler or tunable splitter, enabling a number of potential applications for planar magnonics.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Scalability under a Power Bound using the GREMLINs Framework

With the move towards exascale, system and software developers will have to deal with issues of extreme parallelism. The system properties affected most by the increase in node and core count are the shared resources on node and across the system. The increase in parallelism leads to reduced memory and bandwidth when regarding individual cores. Since power is a limiting factor for supercomputers, and power is not fully utilized in current systems, overprovisioning compute resources is a viable approach to maximized power utilization. To maximize system performance in regard to these changing conditions, it is necessary to understand how resource restrictions impact performance and system behavior. For the purpose of understanding anticipated system properties the GREMLINs Framework was developed. The framework gives the opportunity to add power restrictions, hinder memory properties and introduce faults to study resilience, among others. These features give the opportunity to use current petascale technology to study problems system designers and software developers will have to face when moving towards exascale and beyond. This work describes the initial release of the GREMLINs Framework, developed for this work, and shows how it can be used to study the scaling behavior of proxy applications. These proxy applications represent a selection of HPC workloads important to the scientific community. The proxy applications studied are AMG2013, an algebraic multi-grid linear system solver, CoMD, a classical molecular dynamics proxy application and NEKBONE, an application that uses a high order spectral element method to solve the Navier-Stokes equations. The main interest of these studies lies in analysis regarding their power behavior at scale under a power bound. These findings show how the GREMLINs Framework can help systems and software designers to attain better application performance and can also be used as basis for CPU power balancing tools to use power more efficiently.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Status and future plans for C 3 R&D

C 3 is an opportunity to realize an e + e - collider for the study of the Higgs boson at √s = 250 GeV, with a well defined upgrade path to 550 GeV while staying on the same short facility footprint. C 3 is based on a fundamentally new approach to normal conducting linear accelerators that achieves both high gradient and high efficiency at relatively low cost. Given the advanced state of linear collider designs, the key system that requires technical maturation for C 3 is the main linac. This paper presents the staged approach towards a facility to demonstrate C 3 technology with both Direct (source and main linac) and Parallel (beam delivery, damping ring, ancillary component) R&D. The primary goal of the C 3 Demonstration R&D Plan is to reduce technical and cost risk by building and operating the key components of C 3 at an adequate scale. This R&D plan starts with the engineering design, and demonstration of one cryomodule and will culminate in the construction of a 3 cryomodule linac with pre-production prototypes. This R&D program would also demonstrate the linac rf fundamentals including achievable gradient and gradient stability over a full electron bunch train and breakdown rates. It will also investigate beam dynamics including energy spread, wakefields, and emittance growth. This work will be critical to confirm the suitability of the C 3 beam parameters for the physics reach and detector performance in preparation for a Conceptual Design Report (CDR), as well as for follow-on technology development and industrialization. The C 3 Demonstration R&D Plan will open up significant new scientific and technical opportunities based on development of high-gradient and high-efficiency accelerator technology. It will push this technology to operate both at the GeV scale and mature the technology to be reliable and provide high-brightness electron beams. The timeline for progressing with C 3 technology development will be governed by practical limitations on both the technical progress and resource availability. It consists of four stages: Stage 0) Ongoing fundamental R&D on structure prototypes, damping and vibrations. Stage 1) Advancing the engineering maturity of the design and developing start-to-end simulations including space-charge and wakefield effects. This stage will include testing of strucutres operating at cryogenic temperatures. Beam tests would be performed with high beam current to test full beam loading. Stage 2) Production and testing of the first cryomodule at cryogenic temperatures. This would provide sufficient experimental data to compile a CDR and it is anticipated for Stage 2 to last 3 years and to culminate with the transport of photo-electrons through the first cryomodule. Stage 3) Updates to the engineering design of the cryomodules, production of the second and third cryomodule and their installation. Lower charge and lower emittance beams will be used to investigate emittance growth. The successful full demonstration of the 3 cryomodules to deliver up to a 3 GeV beam and achieve the C 3 five gradient will allow a comprehensive and robust evaluation of the technical design of C 3 as well as mitigate technical, schedule, and cost risks required to proceed with a Technical Design Report (TDR).

radiation hardened magnets↗

High frequency fast wave current drive for DEMO

A steady-state tokamak reactor (SSTR) requires a high efficiency current drive system, from plug to driven mega-amps. RF systems working in the ion-cyclotron range of frequencies (ICRF) have high efficiency from plug to antenna but a limited current drive (CD) efficiency and centrally peaked CD profiles. The latter feature is not adequate for a SSTR where the current should be sufficiently broad to keep the central safety factor (possibly significantly) above 1. In addition, the fact that the fast wave (FW) is evanescent at the edge limits coupling, requiring high voltage operation, which makes the system dependent on plasma edge properties and prone to arcing, reducing its reliability. A possible way to overcome these weaknesses is to operate at higher frequency (10 times or more the cyclotron frequency). The advantages are: (1) The coupling can be much better (waves propagate in vacuum) if the parallel refractive index n{sub ||} is kept below one, (2) The FW group velocity tends to align to the magnetic field, so the power circumnavigates the magnetic axis and can drive off-axis current, (3) Due to the latter property, n{sub ||} can be upshifted along the wave propagation path, allowing low n{sub ||} launch (hence good coupling, large CD efficiency) with ultimately good electron absorption (which requires higher n{sub ||}. Note however that the n{sub ||} upshift is a self-organized feature, that electron absorption is in competition with {alpha}-particle absorption and that uncoupling of the FW from the lower hybrid resonance at the edge requires n{sub ||} slightly above one. The latter possibly counterproductive features might complicate the picture. The different aspects of this potentially attractive off-axis FWCD scheme are discussed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Solar thermophotovoltaics: reshaping the solar spectrum

There has been increasing interest in utilizing solar thermophotovoltaics (STPV) to convert sunlight into electricity, given their potential to exceed the Shockley-Queisser limit. Encouragingly, there have also been several recent demonstrations of improved system-level efficiency as high as 6.2%. Here, we review prior work in the field, with particular emphasis on the role of several key principles in their experimental operation, performance, and reliability. In particular, for the problem of designing selective solar absorbers, we consider the trade-off between solar absorption and thermal losses, particularly radiative and convective mechanisms. For the selective thermal emitters, we consider the tradeoff between emission at critical wavelengths and parasitic losses. Then for the thermophotovoltaic (TPV) diodes, we consider the trade-off between increasing the potential short-circuit current, and maintaining a reasonable opencircuit voltage. This treatment parallels the historic development of the field, but also connects early insights with recent developments in adjacent fields.With these various components connecting in multiple ways, a system-level end-to-end modeling approach is necessary for a comprehensive understanding and appropriate improvement of STPV systems. Our approach will ultimately allow researchers to design STPV systems capable of exceeding recently demonstrated efficiency values.

14 SOLAR ENERGY↗

Assessment of Anisotropic Semiconductor Nanorod and Nanoplatelet Heterostructures with Polarized Emission for Liquid Crystal Display Technology

Semiconductor nanorods can emit linear-polarized light at efficiencies over 80%. Polarization of light in these systems, confirmed through single-rod spectroscopy, can be explained on the basis of the anisotropy of the transition dipole moment and dielectric confinement effects. Here we report emission polarization in macroscopic semiconductor polymer composite films containing CdSe/CdS nanorods and colloidal CdSe nanoplatelets. Anisotropic nanocrystals dispersed in polymer films of poly butyl-co-isobutyl methacrylate (PBiBMA) can be stretched mechanically in order to obtain unidirectionally aligned arrays. A high degree of alignment, corresponding to an orientation factor of 0.87, was achieved and large areas demonstrated polarized emission, with the contrast ratio I-parallel to/I-perpendicular to= 5.6, making these films viable candidates for use in liquid crystal display (LCD) devices. To some surprise, we observed significant optical anisotropy and emission polarization for 2D CdSe nanoplatelets with the electronic structure of quantum wells. The aligned nanorod arrays serve as optical funnels, absorbing unpolarized light and re-emitting light from deep-green to red with quantum efficiencies over 90% and high degree of linear polarization. Our results conclusively demonstrate the benefits of anisotropic nanostructures for LCD backlighting. The polymer films with aligned CdSe/CdS dot-in-rod and rod-in-rod nanostructures show more than 2-fold enhancement of brightness compared to the emitter layers with randomly oriented nanostructures. This effect can be explained as the combination of linearly polarized luminescence and directional emission from individual nanostructures.

liquid crystal display↗

A fast integrated mobility spectrometer for rapid measurement of sub-micrometer aerosol size distribution, Part II: Experimental characterization

A Fast Integrated Mobility Spectrometer (FIMS) with a wide dynamic size range has been developed for rapid aerosol size distribution measurements. The design and model evaluation of the FIMS are presented in the preceding paper (Paper I), and this paper focuses on the experimental characterization of the FIMS. Monodisperse aerosol with diameter ranging from 8 to 600 nm was generated using Differential Mobility Analyzer (DMA), and was measured by the FIMS in parallel with a Condensation Particle Counter (CPC). The mean particle diameter measured by the FIMS is in good agreement with the DMA centroid diameter. Comparison of the particle concentrations measured by the FIMS and CPC indicates the FIMS detection efficiency is essentially 100% for particles with diameters of 8 nm or larger. For particles smaller than 20 nm or larger than 200 nm, FIMS transfer function and resolution can be well represented by the calculated ones based on simulated particle trajectories in the FIMS. For particles between 20 and 200 nm, the FIMS transfer function is boarder than the calculated, likely due to non-ideality of the electric field, including edge effects near the end of the electrode, which are not represented by the 2-D electric field used to simulate particle trajectories.

54 ENVIRONMENTAL SCIENCES↗

A computationally efficient method for full-core conjugate heat transfer modeling of sodium fast reactors

For efficient and accurate temperature predictions of sodium fast reactor structures, a 3-D full-core conjugate heat transfer modeling capability is developed for an advanced system analysis tool, SAM. The hexagon lattice core is modeled with 1-D parallel channels representing the subassembly flow, and 2-D duct walls and inter-assembly gaps. The six sides of the hexagon duct wall and near-wall coolant region are modeled separately to account for different temperatures and heat transfer between coolant flow and each side of the duct wall. The Jacobian Free Newton Krylov (JFNK) solution method is applied to solve the fluid and solid field simultaneously in a fully coupled fashion. The 3-D full-core conjugate heat transfer modeling capability in SAM has been demonstrated by a verification test problem with 7 fuel assemblies in a hexagon lattice layout. In addition, the SAM simulation results are compared with RANS-based CFD simulations. Very good agreements have been achieved between the results of the two approaches.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Drought re-routes soil microbial carbon metabolism towards emission of volatile metabolites in an artificial tropical rainforest

Drought impacts on microbial activity can alter soil carbon fate and lead to the loss of stored carbon to the atmosphere as CO 2 and volatile organic compounds (VOCs). Here we examined drought impacts on carbon allocation by soil microbes in the Biosphere 2 artificial tropical rainforest by tracking 13 C from position-specific 13 C-pyruvate into CO 2 and VOCs in parallel with multi-omics. During drought, efflux of 13 C-enriched acetate, acetone and C 4 H 6 O 2 (diacetyl) increased. These changes represent increased production and buildup of intermediate metabolites driven by decreased carbon cycling efficiency. Simultaneously, 13 C-CO 2 efflux decreased, driven by a decrease in microbial activity. However, the microbial carbon allocation to energy gain relative to biosynthesis was unchanged, signifying maintained energy demand for biosynthesis of VOCs and other drought-stress-induced pathways. Overall, while carbon loss to the atmosphere via CO 2 decreased during drought, carbon loss via efflux of VOCs increased, indicating microbially induced shifts in soil carbon fate.

59 BASIC BIOLOGICAL SCIENCES↗

Design of a high-power, high-gain, 2nd harmonic, 22.848 GHz gyroklystron

In this paper we consider the design of a four-cavity, high-gain K-band gyroklystron experiment for high gradient structure testing. The frequency doubling gyroklystron utilizes a beam voltage of 500 kV and a beam current of 200 A from a magnetron injection gun (MIG) originally designed for a lower-frequency device. The microwave circuit features input and gain cavities in the circular TE{sub 011} mode and penultimate and output cavities that operate at the second harmonic in the TE{sub 021} mode. We investigate the MIG performance and study the behavior of the circuit for different values of perpendicular to parallel velocity ratio (α= V{sub ⊥}/ V{sub z}). This microwave tube is expected to be able to produce at least 20 MW of power in 1μs pulses at a repetition rate of at least 120 Hz. A maximum efficiency of 26% and a large signal gain of 58 dB under zero-drive stable conditions were simulated for a velocity ratio equal to 1.35.

43 PARTICLE ACCELERATORS↗

Validation of transport models using additive flux minimization technique

A new additive flux minimization technique is proposed for carrying out the verification and validation (V and V) of anomalous transport models. In this approach, the plasma profiles are computed in time dependent predictive simulations in which an additional effective diffusivity is varied. The goal is to obtain an optimal match between the computed and experimental profile. This new technique has several advantages over traditional V and V methods for transport models in tokamaks and takes advantage of uncertainty quantification methods developed by the applied math community. As a demonstration of its efficiency, the technique is applied to the hypothesis that the paleoclassical density transport dominates in the plasma edge region in DIII-D tokamak discharges. A simplified version of the paleoclassical model that utilizes the Spitzer resistivity for the parallel neoclassical resistivity and neglects the trapped particle effects is tested in this paper. It is shown that a contribution to density transport, in addition to the paleoclassical density transport, is needed in order to describe the experimental profiles. It is found that more additional diffusivity is needed at the top of the H-mode pedestal, and almost no additional diffusivity is needed at the pedestal bottom. The implementation of this V and V technique uses the FACETS::Core transport solver and the DAKOTA toolkit for design optimization and uncertainty quantification. The FACETS::Core solver is used for advancing the plasma density profiles. The DAKOTA toolkit is used for the optimization of plasma profiles and the computation of the additional diffusivity that is required for the predicted density profile to match the experimental profile.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

A multi-species collisional operator for full-F gyrokinetics

A linearized multi-species collision operator has been developed for an efficient implementation in gyrokinetic codes. This operator satisfies the main expected properties: particle, momentum, and energy conservation, and existence of an H-theorem. A gyrokinetic version is then calculated, which involves derivatives with respect to the gyrocenter position, parallel velocity, and magnetic momentum. An isotropic version in the velocity space can be constructed for the specific problem of trace impurities colliding with a main species. A simpler version that involves derivatives with parallel velocity only has been developed. This reduced version has been implemented in the GYSELA gyrokinetic code, and is shown to comply with particle, momentum, and energy conservation laws. Moreover, the interspecies relaxation rates for momentum and energy agree very well with the theoretical values.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Control of runaway electron energy using externally injected whistler waves

One way of mitigating runaway damage of the plasma-facing components in a tokamak fusion reactor is by limiting the runaway electron energy under a few MeV, while not necessarily reducing the runaway current appreciably. Here in this paper, we describe a physics mechanism by which such momentum space engineering of the runaway distribution can be facilitated by externally injected high-frequency electromagnetic waves such as whistler waves. The drastic impact that wave-induced scattering can have on the runaway energy distribution is fundamentally the result of its ability to control the runaway vortex in the momentum space. The runaway vortex, which is a local circulation of runaways in momentum space, is the outcome of the competition between Coulomb collisions, synchrotron radiation damping, and runaway acceleration by the parallel electric field. By introducing a wave that resonantly interacts with runaways in a particular range of energies which is mildly relativistic, the enhanced scattering would reshape the vortex by cutting off the part that is highly relativistic. The efficiency of resonant scattering accentuates the requirement that the wave amplitude can be small so the power requirement from external wave injection is practical for the mitigation scheme.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Simulation study of CO 2 laser-plasma interactions and self-modulated wakefield acceleration

3D numerical simulations of the interaction of a powerful CO 2 laser with hydrogen jets demonstrating the role of ionization on the characteristics of induced wakes are presented. Simulations using SPACE, a parallel relativistic Particle-in-Cell code, are performed in support of the plasma wakefield accelerator experiments being conducted at the Brookhaven National Laboratory (BNL) Accelerator Test Facility (ATF). A novelty of the SPACE code is its set of efficient atomic physics algorithms that compute ionization and recombination rates on the grid and transfer them to particles. The influence of ionization on the spectrum of pump laser has been studied for a range of gas densities.Simulations reproduce both Stokes and anti-Stokes shifts in the spectrum of pump laser, similar to those observed in experiments in the spectrum of probe. A good agreement has been achieved with the experiments on the effect of variation in gas density on Stokes/anti-Stokes intensity. In addition, self-injection and trapping of electrons into the self-modulated wakes has been observed and analyzed. Here, the experimentally validated code SPACE will be used for predictive simulations to guide future experiments at BNL ATF.

97 MATHEMATICS AND COMPUTING↗

Stability Analysis of Parallel Connected Bidirectional WPT System

This paper presents a stability analysis of parallel-connected bi-directional series-series resonant network wireless power transfer (WPT), optimized for Electric Vehicle (EV) charging and vehicle-to-grid (V2G) applications. The study addresses critical stability challenges in systems integrated with diverse distributed energy resources (DERs), including photovoltaics, fuel cells, wind turbines, energy storage systems, and the AC grid. The stability of such integrated DC grid systems is paramount for ensuring reliable operation, particularly under varying power flow conditions and dynamic interactions between parallel WPT systems. The analysis included system impedance characterization, state-space modeling, and open and closed-loop stability evaluations. The results demonstrated that the integration of a robust control architecture effectively mitigates instability risks and supports scalable, efficient operation. This work underscores the converter's adaptability and its potential for large-scale deployment in wireless EV charging infrastructures and integrated DC grid systems.

Asa, Erdem [ORNL] (ORCID:0000000190884812)↗