Build Times on Modern Heterogeneous Systems: Steps towards a faster future
Explore the source record for details and available documents.
Elliott, James John↗
SEARCH · Search NASA
Results for “faster”
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.
FASTER TENSOR KERNELS VIA RUST-FROM-PYTHON IMPLEMENTATION
Explore the source record for details and available documents.
Harding, Kimberly Ann [Sandia National Laboratorie↗
Technical note: A modified formulation of dynamic energy budget theory for faster computation of biological growth
Abstract. The mass conservation equation in the presence of boundary fluxes and chemical reactions from non-equilibrium thermodynamics is used to derive a modified dynamic energy budget (mDEB) model. Compared to the standard dynamic energy budget (sDEB) model (Kooijman, 2009), this modified formulation does not place the dilution effect in the mobilization kinetics of reserve biomass, and it maintains the partition principle for reserve mobilization dynamics for both linear and non-linear kinetics. Overall, the mDEB model shares most features with the sDEB model. However, for biological growth that requires multiple nutrients, the mDEB model is computationally much more efficient by not requiring numerical iterations for obtaining the specific growth rate. In an example of modeling the growth of Thalassiosira weissflogii in a nitrogen-limiting chemostat, the mDEB model was found to have almost the same accuracy as the sDEB model while requiring almost half of the computing time of the sDEB model. Since the sDEB model has been successfully applied in numerous studies, we believe that the mDEB model can help improve the modeling of biological growth and the associated ecosystem processes in various contexts.
Tang, Jinyun↗
Homomorphic data compression for real time photon correlation analysis
The construction of highly coherent X-ray sources, combined with next-generation detectors that are larger and faster, has enabled new research opportunities across the scientific landscape. Among the techniques that benefit most from these advancements is X-ray photon correlation spectroscopy (XPCS), where faster acquisition unlocks the ability to study faster dynamics within samples. However, faster acquisition on larger detectors also introduces unprecedented challenges for online data processing and offline data storage. Such challenges are particularly prominent for XPCS, where real time analyses require simultaneous calculation of all the previously acquired data in the time series. We present a homomorphic compression scheme to effectively reduce the computational time and memory space required for XPCS analysis. Leveraging similarities in the mathematical expression between a matrix-based compression algorithm and the correlation calculation, our approach allows direct operation on the compressed data without their decompression. The offline compression scheme extends storage capacity by a factor of 40 while preserving key features in the lossy compressed data. Meanwhile, the online compression scheme reduces the computational time to below 1 ms, enabling real time calculation of the correlation functions at kHz framerate. Our demonstration of a homomorphic compression of scientific data provides an effective solution to the big data challenge at coherent light sources. Beyond the example shown in this work, the framework can be extended to facilitate real-time operations directly on a compressed data stream for other techniques.
36 MATERIALS SCIENCE↗
Open‐Source Anaerobic Digestion Modeling Platform, Anaerobic Digestion Model No. 1 Fast (ADM1F)
An open‐source modeling platform, called Anaerobic Digestion Model No. 1 Fast (ADM1F), is introduced to achieve fast and numerically stable simulations of anaerobic digestion processes. ADM1F is compatible with an iPython interface to facilitate model configuration, simulation, data analysis, and visualization. Faster simulations and more stable results are accomplished by implementing an advanced open‐source library of numerical methods called Portable Extensive Toolkit for Scientific Computation (PETSc) to solve the ADM1 system of equations. Leveraging PETSc, ADM1F can consistently complete a steady‐state simulation under 0.2 s, over 99% faster than a benchmark ADM1 model implemented with MATLAB while achieving agreement of model outputs within 1% of those obtained with the benchmark model. For dynamic simulations, however, ADM1F has a computational speed advantage only when the influent characteristics update more frequently than every 4 h. The ability of ADM1F to be useful as a tool to study anaerobic digestion systems is demonstrated through two example implementations of ADM1F: (1) a two‐phase co‐digestion scenario evaluating the impact of the organic loading rate and the substrate composition on reactor performance and stability, and (2) a conventional digester scenario assessing the effectiveness of recovery strategies after disruptions that led to instability. These examples demonstrate how the high simulation speed and the convenience of the iPython interface allow ADM1F to complete complex analyses within minutes, much faster than computational strategies currently reported in the literature.
anaerobic co-digestion↗
Evaporation from thin porous Coatings: Pore size effects and predictive equation for homogeneous coatings
Evaporation of small water droplets on solids is hindered because surface tension pulls the droplet into a spherical cap that has a small perimeter. Our solution is to coat a solid with a very thin, porous layer into which the droplet flows to create a large-area disk with concomitant high rate of evaporation. We investigate evaporation by varying factors that have not been previously considered: pore size and distribution, contact angle, temperature, and relative humidity (RH). A larger pore size resulted in faster evaporation, which we explain through faster transport within the coating. Even faster evaporation occurred for a bilayer structure with small particles on the air side and larger particles on the solid side. Further, the water advancing contact angle had an insignificant effect in the range from < 10° through to 60°. Our results for different pore sizes, temperature, humidity, and contact angle all collapse onto a single curve when appropriately normalized. This validates an equation that can be used for the evaporation from a homogeneous coating that depends only one empirical factor and the droplet volume. Since the volume is often user-controlled, we envisage that this equation can be used to predict evaporation and guide design of fast-drying coatings.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
Ni coarsening under humid atmosphere in the electrode of solid oxide cells: A combined study of density-functional theory and phase-field modeling
Ni coarsening in solid oxide cell (SOC) electrodes is known to be significantly faster under a humid atmosphere. The underlying mechanisms, though, have not been fully understood. In this work, we examine the surface diffusion of Ni(OH) x by a combination of density-functional theory and phase-field modeling. Here, density-functional theory is used to evaluate the adsorption and diffusion of the Ni(OH) x species on Ni (111) surface, and the results are used to obtain the effective surface diffusivity of Ni versus steam and hydrogen gas partial pressures. This diffusivity is used as an input for a phase-field model to investigate the Ni coarsening in SOC electrodes. It is found that Ni(OH) x formed through chemical reaction cannot accelerate coarsening unless an extremely high steam to hydrogen ratio is reached. However, if Ni(OH) x is formed through electrochemical reactions near triple phase boundaries (TPBs), surface diffusion of Ni(OH) x may cause faster Ni coarsening in fuel cell mode under a large overpotential. Specifically, surface diffusion of Ni(OH) may be comparable to or faster than that of Ni under an overpotential that is large but still possible under fuel cell operating conditions.
25 ENERGY STORAGE↗
Effect of Lithium Doping on MgO Hydroxylation and Carbonation
Recovery of magnesium from brines can potentially be used to source MgO (periclase) as a CO 2 sorbent or for Mg-based cements. However, it is not clear how common impurities in brines, such as lithium, affect the resulting MgO reactivity. Here, to test the effect of lithium incorporation on MgO reactivity for hydration and carbonation, we combined computational simulations with experiments. Experimentally altered (Mg,Li)O with a low dopant concentration (0.012 ± 0.002% w/w Li) was characterized using synchrotron-based X-ray scattering and high-resolution electron microscopy to measure reaction layer formation on (Mg,Li)O. Single-crystal X-ray diffraction analysis of (Mg,Li)O demonstrates that the incorporation of lithium leads to the formation of oxygen vacancies. The presence of vacancies is likely causing faster hydroxylation rates as predicted by ab initio molecular dynamics simulations. However, the faster hydroxylation rates likely lead to faster passivation of the surface because we observe thinner reaction layers on (Mg,Li)O samples both over short time periods (30 days) and over long time periods (28 years). After 28 years, the reaction layer on the (Mg,Li)O sample was less than one-third of the thickness of that of the pure MgO sample. In addition, over 30 days, reaction layers on (Mg,Li)O samples primarily formed at steps rather than on terraces, in contrast to our previous observations on MgO. Based on our results, naturally occurring impurities in MgO modify its reactivity even at very low concentrations and need to be considered for accurate reaction rate prediction for application of MgO as a CO 2 sorbent or in cements.
MgO↗
Optimization of a Mixed Fleet of Aerial Drones for Medical Supplies: A Case Study of Blood Delivery Logistics
Aerial drones have emerged as an innovative solution for faster transportation of time-sensitive items (e.g., emergency medical supplies), potentially reducing the transmission of contagious diseases and enhancing healthcare availability through contactless autonomous delivery. We study fleet sizing and efficient scheduling of a mixed fleet of drones for delivering time-sensitive medical items having distinct release and due times to minimize the required fleet size and fleet composition, the required number of additional batteries, and the total energy consumption. We continuously track the remaining battery energy of drones to determine the optimal timing for battery replacement, rather than replacing the battery at each node. Using actual drone flight test data, we employed a machine learning (ML) method to estimate the energy consumption of different drone types during flight segments for different operating parameters. We present a novel mixed-integer programming model to efficiently formulate the problem that integrates the estimated energy consumption functions from ML. We propose a new greedy heuristic (GH) algorithm and a customized genetic algorithm (GA) for solving large-scale instances of this problem faster. Results demonstrate that the GH algorithm is substantially faster than the accelerated CPLEX and the GA, while sacrificing the solution quality by a small amount. Results based on an actual blood sample delivery case study from Pendleton, Oregon, United States, show that using a mixed fleet of drones reduces the total cost and total energy consumption up to 18.18% and 28.7%, respectively, compared to using a homogeneous fleet.
29 - ENERGY PLANNING, POLICY AND ECONOMY↗
A graphics processing unit accelerated sparse direct solver and preconditioner with block low rank compression
We present the GPU implementation efforts and challenges of the sparse solver package STRUMPACK. The code is made publicly available on github with a permissive BSD license. STRUMPACK implements an approximate multifrontal solver, a sparse LU factorization which makes use of compression methods to accelerate time to solution and reduce memory usage. Multiple compression schemes based on rank-structured and hierarchical matrix approximations are supported, including hierarchically semi-separable, hierarchically off-diagonal butterfly, and block low rank. Here, in this paper, we present the GPU implementation of the block low rank (BLR) compression method within a multifrontal solver. Our GPU implementation relies on highly optimized vendor libraries such as cuBLAS and cuSOLVER for NVIDIA GPUs, rocBLAS and rocSOLVER for AMD GPUs and the Intel oneAPI Math Kernel Library (oneMKL) for Intel GPUs. Additionally, we rely on external open source libraries such as SLATE (Software for Linear Algebra Targeting Exascale), MAGMA (Matrix Algebra on GPU and Multi-core Architectures), and KBLAS (KAUST BLAS). SLATE is used as a GPU-capable ScaLAPACK replacement. From MAGMA we use variable sized batched dense linear algebra operations such as GEMM, TRSM and LU with partial pivoting. KBLAS provides efficient (batched) low rank matrix compression for NVIDIA GPUs using an adaptive randomized sampling scheme. The resulting sparse solver and preconditioner runs on NVIDIA, AMD and Intel GPUs. Interfaces are available from PETSc, Trilinos and MFEM, or the solver can be used directly in user code. We report results for a range of benchmark applications, using the Perlmutter system from NERSC, Frontier from ORNL, and Aurora from ALCF. For a high frequency wave equation on a regular mesh, using 32 Perlmutter compute nodes, the factorization phase of the exact GPU solver is about 6.5× faster compared to the CPU-only solver. The BLR-enabled GPU solver is about 13.8× faster than the CPU exact solver. For a collection of SuiteSparse matrices, the STRUMPACK exact factorization on a single GPU is on average 1.9× faster than NVIDIA’s cuDSS solver.
97 MATHEMATICS AND COMPUTING↗
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↗
Accelerating Neutrino Event Generation in MARLEY Using CUDA-Based RNG and GPU Parallelization
MARLEY is a simulation tool that helps scientists study how low-energy neutrinos interact with matter. To work properly, MARLEY uses random numbers thousands of times in each simulation. These random numbers are important for modeling things like how neutrinos collide with atoms and what particles they produce. Right now, MARLEY runs on a regular computer processor (CPU) and uses a built-in random number generator called the Mersenne Twister. This setup works, but it can be slow, especially when trying to simulate many events. This research focuses on making MARLEY run faster by moving the random number generation and some of the repetitive calculations from the CPU to a graphics processing unit (GPU), which can handle many tasks at the same time. We use CUDA (a tool for programming NVIDIA GPUs) and cuRAND (a GPU-based random number library) to test faster alternatives to the current random number system. We compare different GPU-based generators, like curand_mtgp32, xorwow, and philox, to see which ones are the quickest and still give reliable results. Early tests show that using the GPU can make MARLEY simulations much faster. This project not only helps improve current simulation performance but also moves closer to a full simulation chain where all stages can run on modern GPU hardware.
Dunkley, Kimieka [Florida A-M]↗
Exploring the Effects of Varying Pre-Chamber Geometry in a Heavy-Duty Natural Gas Optical Engine under Dilution Conditions
Pre-chamber combustion is an advanced ignition strategy that has been shown to enhance spark ignition (SI) combustion stability in natural gas (NG) engines by providing distributed ignition sites from turbulent jets and enhancing main-chamber turbulence. Pre-chamber combustion has been proven especially advantageous compared to SI in ultra-lean and dilute operating conditions. This work involves experimental investigation of the effects of varying passive pre-chamber nozzle configuration on pre-chamber and main chamber combustion under simulated exhaust gas recirculation (EGR) dilution (0 and 20%) conditions in a heavy-duty, single-cylinder, optically accessible NG engine at stoichiometric fuel-air ratio. Pre-chamber nozzle configurations include four pre-chambers with constant nozzle area to pre-chamber volume ratio (A/V) with different nozzle sizes and orientations and one configuration with larger nozzles. The optical engine is operated in a skip-fire sequence consisting of 18 motored cycles followed by two consecutive fired cycles to elucidate the effect of combustion residuals (internal EGR) on combustion evolution. Pressure-based diagnostics are used to monitor pre-chamber and subsequent main chamber combustion, and optical diagnostics include high-speed OH* chemiluminescence to visualize the development of pre-chamber jets and the resulting ignition of the main chamber charge. Heat release analysis of the in-cylinder pressure data indicates that a faster pre-chamber pressure rise does not always translate into faster main-chamber combustion. The pre-chamber with the smallest nozzle diameter produced the highest pre-chamber pressure rise and fastest combustion under non-diluted conditions. However, dilution delays the main chamber ignition for smaller nozzles despite a comparable rise in pre-chamber pressure compared to configurations with larger nozzles. This effect is more pronounced for cycles with in-cylinder combustion residuals in addition to external dilution. Additionally, it was observed that pre-chambers with swirling nozzles have a faster pressure rise in the pre-chamber and main chamber under dilute conditions. Optical diagnostics suggest that the main reason for the delay between the pre-chamber pressure rise and main-chamber combustion lies in jet quenching and delayed re-ignition, which can even lead to misfire if jets emitted from small nozzles combined with dilution fail to re-ignite.
Dhotre, Akash [University Of Minnesota-Twin Cities↗
Novel Mito-Nuclear Combinations Facilitate the Global Invasion of a Major Agricultural Crop Pest
A fundamental understanding of the underlying mechanisms involved in biological invasions is crucial to developing effective risk assessment and control measures against invasive species. The fall armyworm (FAW), Spodoptera frugiperda, is a highly invasive pest that has rapidly spread from its native Americas into much of the Eastern Hemisphere, with a highly homogeneous nuclear genetic background. However, the exact mechanism behind its rapid introduction and propagation remains unclear. Here, a systematic investigation is conducted into the population dynamics of FAW in China from 2019 to 2021 and found that FAW individuals carrying “rice” mitochondria (FAW-mR) are more prevalent (>98%) than that with “corn” mitochondria (FAW-mC) at the initial stage of the invasion and in newly-occupied non-overwintering areas. Further fitness experiments show that the two hybrid-strains of FAW exhibit different adaptions in the new environment in China, and this may have been facilitated by amino acid changes in mitochondrial-encoded proteins. FAW-mR used increases energy metabolism, faster wing-beat frequencies, and lower wing loadings to drive greater flight performance and subsequent rapid colonization of new habitats. In contrast, FAW-mC individuals adapt with more relaxed mitochondria and shuttle energetics into maternal investment, observed as faster development rate and higher fecundity. The presence of two different mitochondria types within FAW has the potential to significantly expand the range of damage and enhance competitive advantage. Overall, the study describes a novel invasion mechanism displayed by the FAW population that facilitates its expansion and establishment in new environments.
60 APPLIED LIFE SCIENCES↗
Utilization of Data Augmentation Techniques in Automated Inspection Systems for Defect Detection in Metals With Limited Data
Accurate identification of defects on metal surfaces is of great interest to many industry sectors, such as the automotive and aerospace industries. In contrast to conventional manual inspection techniques, recent automated inspection systems employ deep learning models trained to detect defects rapidly and precisely. The development of these models often requires a substantial image dataset to acquire adequate knowledge of defect features and enhance their predictive accuracy. When data is limited, augmentation techniques are often used to improve the precision and accuracy of defect detection systems. This study examined the prediction performance of two object detection models, namely Faster Region‐based Convolutional Neural Network (Faster R‐CNN) and You Only Look Once version 8 (YOLOv8), to identify dent defects in limited images of cast iron cylinder head surfaces. The original image set contains 46 images with 563 dents. To overcome limited data availability, common image augmentation techniques along with a copy‐paste method were applied. Results show that standard augmentation improved YOLOv8 accuracy by 8.00% and average precision (AP) by 3.00%. On the other hand, the copy‐paste technique achieved a 20.00% increase in accuracy and a 1% increase in AP with just 200 synthetic dents. Furthermore, these results provide support for using the copy‐paste augmentation strategy to enhance defect detection performance, with a limited dataset, contributing to more accurate defect identification in remanufacturing processes.
36 MATERIALS SCIENCE↗
Unraveling Electronic Effects that Influence the Bimolecular Chemistry of Fluorinated Criegee Intermediates
Ozonolysis is important for the removal of unsaturated volatile organic compounds emitted into the Earth's atmosphere. The alkene + ozone reaction (alkene ozonolysis) forms a carbonyl oxide – a Criegee intermediate (CI). CIs are interesting from an atmospheric chemistry perspective. They are implicated in enhancing the oxidizing capacity of the troposphere and in secondary organic aerosol formation. CIs may participate in bimolecular reactions with trace atmospheric vapors like water and methanol. Here, the chemistry of the two isomers of CFHOO and of CF 3 CHOO with water, water dimer, and methanol is explored. The present study confirms that the syn-conformer is the minimum energy structure for both CIs and that the reactions of anti-CFHOO display slightly lower activation energies cf. syn-CFHOO, thus promoting faster reaction. For both syn- and anti-CFHOO, the energy barrier in the pathway to adding water, water dimer, or methanol is lower—and the overall reaction exoergicity greater—than for the corresponding reactions involving the nonfluorinated analogue, CH 2 OO. Here, the corresponding bimolecular reaction rates are thus predicted to be faster for CFHOO than CH 2 OO. Similar reactivity trends are found when comparing the same reactions of syn-CF 3 CHOO, cf. syn-CH 3 CHOO, but not when comparing the energetics of the reactions involving anti-CF 3 CHOO cf. anti-CH 3 CHOO.
Guidry, Lily M. [University of Louisiana at Lafaye↗
Electro-carburization of mild steel through two-step reaction in molten-salt electrolyte
The rapid carburization of mild steel is demonstrated via the electrochemical deposition of solid carbon, followed by thermochemical iron-carbon (Fe-C) reaction using a molten lithium-potassium carbonate (LiKCO 3 ) salt electrolyte and carbon dioxide as the carbon source. Depth-dependent microstructural and compositional changes in the steel are shown to be controlled by the applied cell voltage and reaction temperatures. Reaction temperatures above the Fe-C eutectoid temperature promote faster carburization due to carbon dissolution into austenite. The carbon deposition rate is regulated by the applied cell voltage: At - 2.4 V, carbon deposition occurs much faster than its diffusion into the steel, resulting in the accumulation of a carbon-rich layer exceeding 1 mm, which does not diffuse into the steel. In contrast, at - 1.8 V, all deposited carbon effectively diffuses into the steel, leaving a minimal external deposit. In both cases, carbon ingress forms a carbide-rich surface layer and precipitates additional carbides along the grain boundaries within the steel electrode. This method offers a potential approach to increase the carbon content of mild or low-carbon steel feedstocks, such as scrap or direct reduced iron, for industrial applications.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
Multi-frequency signatures of space-leader evolution in negative cloud-to-ground lightning stepped leaders
In this study, we examined 364 space leaders in 18 negative natural cloud-to-ground lightning strokes whose stepped leaders created new channels to ground. All strokes were captured on ultra-high-speed video cameras operating at frame rates ranging from 400k to 783k frames per second. Additionally, broadband electromagnetic field measurements were available for a subset of these strokes. The median space leader inception-to-attachment-point length and retrograde propagation speed towards the pre-existing leader channel (PELC) were 8.2 m and 4.0 x 10 6 m/s, respectively. Space leader lengths were longer and retrograde propagation speeds faster for return strokes with higher peak currents. This is likely due to the relative proximity of space leader inception points to the PELC, which makes the electric field produced by the PELC line charge density one of the primary factors in determining space leader characteristics. Space leader characteristics were weakly related to their inception altitude. We observed bursts of very high frequency (VHF) emissions preceding, by around 0.5 – 1 μs, electric field leader-step pulses; visible-frequency-range luminosity pulses started during the step pulses. The median downward leader propagation speed for all 18 strokes was 4.3 x 10 5 m/s; leader propagation speeds were generally faster for return strokes with higher peak currents. Also, leaders appeared to accelerate (on their way to ground) at altitudes lower than about 200 and 1000 m above ground level for strokes in the 10 – 60 and 84 – 228 kA peak current ranges, respectively.
54 ENVIRONMENTAL SCIENCES↗