Search NASA⌕ Search

SEARCH · Search NASA

Results for “Simulation”

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 1,531 records · Page 85
All sourcesNASA NTRSDOE OSTI
DOE OSTI2019

Shattered pellet injection simulations with NIMROD

Optimal strategies for disruption mitigation benefit from the understanding of details both spatially and temporally. Beyond the assessment of the efficacy of a particular proposed Disruption Mitigation System (DMS), ITER's longevity will require accounting of both mitigated and unmitigated disruptions. Accurate models and validated simulations that detail multiple ITER scenarios with mitigated and unmitigated disruptions are essential for accurate estimates of load damage. The primary candidate for ITER's DMS is Shattered Pellet Injection (SPI); its efficacy must be evaluated within the next several years. To perform critical time dependent 3-D nonlinear simulations, we have developed a particle based SPI model in the NIMROD code coupled to its modified single fluid equations with impurity and radiation [Izzo, Nucl. Fusion 46(5), 541 (2006)]. SPI validation simulations of the thermal quench and comparisons to DIII-D impurity scan experiments [Shiraki et al., Phys. Plasmas 23(6), 062516 (2016)] are presented. We also present an initial ITER Q = 10 pure neon SPI simulation and compare it with the DIII-D SPI simulations. NIMROD SPI simulations demonstrate that the ablating fragment drives strong parallel flows that transport the impurities and governs the thermal quench. Analysis of SPI simulations shows that the mixed deuterium/neon SPI results in a more benign thermal quench due to the enhanced transport caused by the additional deuterium. These results suggest that an optimal pellet mixture exists for the SPI system.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Considerations for the modeling of the laser ablation region of ICF targets with Lagrangian simulations

Recently, much effort has been dedicated to the improvement of models and modeling choices utilized in radiation hydrodynamic simulations of direct drive inertial confinement fusion experiments in an effort to improve their predictive capability. In this paper, we consider the choice in mesh for the simulation of the laser ablation of a direct-drive-like target and compare Lagrangian simulations with various mesh zoning choices with Eulerian simulations with fixed resolution in the laser energy deposition region. Using these simulations, we demonstrate how errors in ablation pressure, laser deposition rate, shock speed, and density profile arise from insufficient zoning following from the conservation of mass of Lagrangian zones. These considerations place stringent requirements on the initial t = 0 zoning in the solid density shell for simulations aiming at resolving the ablation and laser absorption region. However, with sufficiently fine zoning in the t = 0 shell, agreement with Eulerian simulations and analytic scaling laws can be recovered.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

Efficient fully-coherent quantum signal processing algorithms for real-time dynamics simulation

Simulating the unitary dynamics of a quantum system is a fundamental problem of quantum mechanics, in which quantum computers are believed to have significant advantage over their classical counterparts. One prominent such instance is the simulation of electronic dynamics, which plays an essential role in chemical reactions, non-equilibrium dynamics, and material design. These systems are time-dependent, which requires that the corresponding simulation algorithm can be successfully concatenated with itself over different time intervals to reproduce the overall coherent quantum dynamics of the system. In this paper, we quantify such simulation algorithms by the property of being fully-coherent: the algorithm succeeds with arbitrarily high success probability 1 − δ while only requiring a single copy of the initial state. Here we subsequently develop fully-coherent simulation algorithms based on quantum signal processing (QSP), including a novel algorithm that circumvents the use of amplitude amplification while also achieving a query complexity additive in time t, ln(1/δ), and ln(1/ϵ) for error tolerance ϵ: $Θ‖\mathscr{H}‖|t|+ln(1/ϵ)+ln(1/δ)$. Furthermore, we numerically analyze these algorithms by applying them to the simulation of the spin dynamics of the Heisenberg model and the correlated electronic dynamics of an H2 molecule. Since any electronic Hamiltonian can be mapped to a spin Hamiltonian, our algorithm can efficiently simulate time-dependent ab initio electronic dynamics in the circuit model of quantum computation. Accordingly, it is also our hope that the present work serves as a bridge between QSP-based quantum algorithms and chemical dynamics, stimulating a cross-fertilization between these exciting fields.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Analysis of single-mode Richtmyer–Meshkov instability using high-order incompressible vorticity-streamfunction and shock-capturing simulations

In this study, two- and three-dimensional simulation results obtained using a new high-order incompressible, variable-density vorticity–streamfunction (VS) method and data from previous ninth-order weighted essentially nonoscillatory (WENO) shock-capturing simulations are used to investigate the nonlinear dynamics of single-mode Richtmyer–Meshkov instability using a model of a Mach 1.3 air(acetone)/SF6 shock tube experiment. A comparison of the density fields from both simulations with the experimental images demonstrates very good agreement in the large-scale structure with both methods but differences in the small-scale structure. The WENO method captures the small-scale disordered structure observed in the experiment, while the VS method partially captures such structure and yields a strong rotating core. The perturbation amplitude growth from the simulations generally agrees well with the experiment. The simulation bubble and spike amplitudes agree well at early times. At later times, the WENO bubble amplitude is smaller than the VS amplitude and vice versa for the spike amplitude. The predictions of nonlinear single-mode instability growth models are shown to agree with the simulation amplitudes at early-to-intermediate times but underpredict the amplitudes at later times in the nonlinear regime. Visualizations of the mass fraction and enstrophy isosurfaces, velocity and vorticity fields, and baroclinic vorticity production and vortex stretching terms from the three-dimensional simulations indicate that, with the exception of the small-scale structure within the rollups, the VS and WENO results are in good agreement.

42 ENGINEERING↗
DOE OSTI2024

Data Science Shows that Entropy Correlates with Accelerated Zeolite Crystallization in Monte Carlo Simulations

We have performed a data science study of Monte Carlo simulation trajectories to understand factors that can accelerate formation of zeolite nanoporous crystals, a process that can take days or even weeks. In previous work, Monte Carlo simulations predicted and experiments confirmed that using a secondary organic structure-directing agent (OSDA) accelerates crystallization of all-silica LTA zeolite, with experiments finding a three-fold speedup [PCCP 24, 142-148 (2022)]. However, it remains unclear what physical factors cause the speed-up. Here, we apply data science to analyze the simulation trajectories to discover what drives accelerated zeolite crystallization in Monte Carlo going from a one-OSDA synthesis (1OSDA) to a two-OSDA version (2OSDA). We encoded simulation snapshots using the Smooth Overlap of Atomic Positions approach, which represents all 2- and 3-body correlations within a given cutoff distance. Principal component analyses failed to discriminate datasets of structures from 1OSDA and 2OSDA simulations, while the Support Vector Machine (SVM) approach succeeded at classifying such structures with an area-under-curve (AUC) score of 0.99 (where AUC = 1 is a perfect classification) with all 3-body correlations, and as high as 0.94 with only 2-body correlations. SVM decision functions reveal relatively broad / narrow histograms for 1OSDA / 2OSDA datasets, suggesting that the two simulations differ strongly in information heterogeneity. Informed by these results, we performed pair (2-body) entropy calculations during crystallization, resulting in entropy differences that semi-quantitatively account for the speedup observed in the previous Monte Carlo simulations. We conclude that altering synthesis conditions in ways that substantially changes the entropy of labile silica networks may accelerate zeolite crystallization, and we discuss possible approaches for achieving such acceleration.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2026

Comparing ICME simulations with scaled laboratory experiment

In stellar physics and astrophysics, numerical simulations and laboratory experiments are often compared to observational data to support their representation of the real world. However, there is also merit in comparing numerical simulations to properly scaled experiments, especially when the experiment and the simulation are both emulating the solar phenomena. Confirming the credibility of scaled experiments and their scaling with well-validated models is important to expand our knowledge of the associated physical phenomena. This is significant because experiments and simulations can be performed frequently, whereas observations may be limited by location, field of view, and missing data. In this work, we use the Alfvén Wave Solar atmosphere Model, a well-validated magnetohydrodynamic model, to simulate an interplanetary coronal mass ejection (ICME) and compare it to an experiment which provides a scaled analog to a physical ICME. The experiment was performed on the Big Red Ball facility and scaled using dimensionless parameters such as plasma β and magnetosonic Mach number to reproduce the main structure of an ICME. We compare the model-simulated temperature, density, and magnetic field to those from the experiment, as well as the scaling parameters used in the experiment, to those calculated from the simulation. This comparison is performed to further justify the scaling arguments made by the experiment. Additionally, the comparison would lead to the development of stronger scaling arguments for future experiments.

Bryant, K. [University of Michigan, Ann Arbor, MI ↗
DOE OSTI2019

Kinetic simulation of a low-pressure helium discharge with comparison to experimental measurements

Modern computational validation efforts rely on comparison of known experimental quantities such as current, voltage, particle densities, and other plasma properties with the same values determined through simulation. A discrete photon approach for radiation transport was recently incorporated into a particle-in-cell/direct simulation Monte Carlo code. As a result, spatially and temporally resolved synthetic spectra may be generated even for non-equilibrium plasmas. The generation of this synthetic spectra lends itself to potentially new validation opportunities. In this work, initial comparisons of synthetic spectra are made with experimentally gathered optical emission spectroscopy. A custom test apparatus was constructed that contains a 0.5 cm gap distance parallel plane discharge in ultra high purity helium gas (99.9999%) at a pressure of 75 Torr. Plasma generation is initiated with the application of a fast rise-time, 100 ns full-width half maximum, 2.0 kV voltage pulse. Transient electrical diagnostics are captured along with time-resolved emission spectra. A one-dimensional simulation is run under the same conditions and compared against the experiment to determine if sufficient physics are included to model the discharge. To sync the current measurements from experiment and simulation, significant effort was undertaken to understand the kinetic scheme required to reproduce the observed features. Additionally, the role of the helium molecule excimer emission and atomic helium resonance emission on photocurrent from the cathode are studied to understand which effect dominates photo-feedback processes. Results indicate that during discharge development, atomic helium resonance emission dominates the photo-flux at the cathode even though it is strongly self-absorbed. A comparison between the experiment and simulation demonstrates that the simulation reproduces observed features in the experimental discharge current waveform. Here, the synthesized spectra from the kinetic method produces more favorable agreement with the experimental data than a simple local thermodynamic equilibrium calculation and is a first step towards using spectra generated from a kinetic method in validation procedures. The results of this study produced a detailed compilation of important helium plasma chemistry reactions for simulating transient helium plasma discharges.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

FINETUNA: fine-tuning accelerated molecular simulations

Abstract Progress towards the energy breakthroughs needed to combat climate change can be significantly accelerated through the efficient simulation of atomistic systems. However, simulation techniques based on first principles, such as density functional theory (DFT), are limited in their practical use due to their high computational expense. Machine learning approaches have the potential to approximate DFT in a computationally efficient manner, which could dramatically increase the impact of computational simulations on real-world problems. However, they are limited by their accuracy and the cost of generating labeled data. Here, we present an online active learning framework for accelerating the simulation of atomic systems efficiently and accurately by incorporating prior physical information learned by large-scale pre-trained graph neural network models from the Open Catalyst Project. Accelerating these simulations enables useful data to be generated more cheaply, allowing better models to be trained and more atomistic systems to be screened. We also present a method of comparing local optimization techniques on the basis of both their speed and accuracy. Experiments on 30 benchmark adsorbate-catalyst systems show that our method of transfer learning to incorporate prior information from pre-trained models accelerates simulations by reducing the number of DFT calculations by 91%, while meeting an accuracy threshold of 0.02 eV 93% of the time. Finally, we demonstrate a technique for leveraging the interactive functionality built in to Vienna ab initio Simulation Package (VASP) to efficiently compute single point calculations within our online active learning framework without the significant startup costs. This allows VASP to work in tandem with our framework while requiring 75% fewer self-consistent cycles than conventional single point calculations. The online active learning implementation, and examples using the VASP interactive code, are available in the open source FINETUNA package on Github.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

The Concentration–Mass relation of massive, dynamically relaxed galaxy clusters: agreement between observations and ΛCDM simulations

ABSTRACT The relationship linking a galaxy cluster’s total mass with the concentration of its mass profile and its redshift is a fundamental prediction of the Cold Dark Matter (CDM) paradigm of cosmic structure formation. However, confronting those predictions with observations is complicated by the fact that simulated clusters are not representative of observed samples where detailed mass profile constraints are possible. In this work, we calculate the symmetry-peakiness-alignment (SPA) morphology metrics for maps of X-ray emissivity from the three hundred project hydrodynamical simulations of galaxy clusters at four redshifts, and thereby select a sample of morphologically relaxed, simulated clusters, using observational criteria. These clusters have on average earlier formation times than the full sample, confirming that they are both morphologically and dynamically more relaxed than typical. We constrain the concentration–mass–redshift relation of both the relaxed and complete sample of simulated clusters, assuming power-law dependences on mass (κm) and 1 + z (κζ), finding κm = −0.12 ± 0.07 and κζ = −0.27 ± 0.19 for the relaxed subsample. From an equivalently selected sample of massive, relaxed clusters observed with Chandra, we find κm = −0.12 ± 0.08 and κζ = −0.48 ± 0.19, in good agreement with the simulation predictions. The simulated and observed samples also agree well on the average concentration at a pivot mass and redshift providing further validation of the ΛCDM paradigm in the properties of the largest gravitationally collapsed structures observed. This also represents the first clear detection of decreasing concentration with redshift, a longstanding prediction of simulations in data.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Agora : Multicomponent simulation for cross-survey science

ABSTRACT Next-generation cosmological surveys will observe large portions of the sky, with significant overlap between them. Multi-wavelength observations will enable us to analyse the same large-scale structure from different angles using a variety of tracers and astrophysical effects. The complex interplay between these observables calls for a model that can accurately and coherently describe their collective behaviour, posing a challenge that can only be met through the use of simulations. In this work, a suite of simulated extragalactic skies is presented, including maps and/or catalogues of cosmic microwave background (CMB) lensing, thermal and kinetic Sunyaev–Zel’dovich (tSZ/kSZ) effects, cosmic infrared background (CIB), radio sources, galaxy overdensity and galaxy weak lensing. Each of these probes is implemented in the lightcone using halo catalogues and/or particles from the Multidark-Planck2 ($\small {MDPL2}$) N-body simulation, and the modelling is calibrated using hydrodynamic simulations and publicly available data. The auto and cross-spectra of the individual probes, as well as the cross-spectra between the observables, are shown to be consistent with theoretical models and measurements from data. The simulation is shown to have a wide range of applications, including forecasting, pipeline testing, and evaluating astrophysical biases in cross-correlation measurements. It is further demonstrated that the simulation products produced in this work have sufficient accuracy to recover the input cosmology when subjected to a full cosmological analysis and are ready for application in real-world analyses for ongoing and future surveys. The simulation products presented in this work can be accessed at: https://yomori.github.io/agora/index.html.

Astronomy & Astrophysics↗
DOE OSTI2018

DES Y1 Results: validating cosmological parameter estimation using simulated Dark Energy Surveys

We use mock galaxy survey simulations designed to resemble the Dark Energy Survey Year 1 (DES Y1) data to validate and inform cosmological parameter estimation. When similar analysis tools are applied to both simulations and real survey data, they provide powerful validation tests of the DES Y1 cosmological analyses presented in companion papers. We use two suites of galaxy simulations produced using different methods, which therefore provide independent tests of our cosmological parameter inference. The cosmological analysis we aim to validate is presented in DES Collaboration et al. (2017) and uses angular two-point correlation functions of galaxy number counts and weak lensing shear, as well as their cross-correlation, in multiple redshift bins. While our constraints depend on the specific set of simulated realisations available, for both suites of simulations we find that the input cosmology is consistent with the combined constraints from multiple simulated DES Y1 realizations in the $\Omega_m-\sigma_8$ plane. For one of the suites, we are able to show with high confidence that any biases in the inferred $S_8=\sigma_8(\Omega_m/0.3)^{0.5}$ and $\Omega_m$ are smaller than the DES Y1 $1-\sigma$ uncertainties. For the other suite, for which we have fewer realizations, we are unable to be this conclusive; we infer a roughly 70% probability that systematic biases in the recovered $\Omega_m$ and $S_8$ are sub-dominant to the DES Y1 uncertainty. Furthermore, as cosmological analyses of this kind become increasingly more precise, validation of parameter inference using survey simulations will be essential to demonstrate robustness.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Quantum dynamics simulation of the advection-diffusion equation

The advection-diffusion equation is simulated via several quantum algorithms. Three formulations are considered: (1) Trotterization, (2) variational quantum time evolution (VarQTE), and (3) adaptive variational quantum dynamics simulation (AVQDS). These schemes were originally developed for the Hamiltonian simulation of many-body quantum systems. The finite-difference discretized operator of the transport equation is formulated as a Hamiltonian and solved without the need for ancillary qubits. Computations are conducted on a quantum simulator (IBM Qiskit Aer) and a superconducting quantum hardware (IBM Fez). The former emulates the latter without the noise. The actual hardware implementation experiences significant noise. The results of the quantum simulator are compared with data from direct numerical simulation (DNS) with infidelities of the order 10 −5 . In the quantum simulator, Trotterization is observed to have the lowest infidelity and is suitable for fault-tolerant computation. The AVQDS algorithm requires the lowest gate count and circuit depth. The VarQTE algorithm is the next best in terms of gate counts, but the number of its optimization variables is directly proportional to the number of qubits. Due to current hardware limitations, Trotterization cannot be implemented, as it has an overwhelmingly large number of operations. Meanwhile, AVQDS and VarQTE can be executed at the hardware level. These algorithms present a new paradigm for computational transport phenomena on quantum computers.

Alipanah, Hirad [Univ. of Pittsburgh, PA (United S↗
DOE OSTI2021

Deep learning simulations of the microwave sky

Here we present 500 high-resolution, full-sky millimeter-wave deep learning (DL) simulations that include lensed CMB maps and correlated foreground components. We find that these MillimeterDL simulations can reproduce a wide range of non-Gaussian summary statistics matching the input training simulations, while only being optimized to match the power spectra. The procedure we develop in this work enables the capability to mass produce independent full-sky realizations from a single expensive full-sky simulation, when ordinarily the latter would not provide enough training data. We also circumvent a common limitation of high-resolution DL simulations that they be confined to small sky areas, often due to memory or GPU issues; we do this by developing a “stitching” procedure that can recover the large-scale, high-order statistics and avoid discontinuities or repeated features in the maps. In addition, since our network takes as input a full-sky lensing convergence map, it can in principle take a full-sky lensing convergence map from any large-scale structure (LSS) simulation and generate the corresponding lensed CMB and correlated foreground components at millimeter wavelengths; this is especially useful in the current era of combining results from both CMB and LSS surveys, which require a common set of simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2017

First-principles equation-of-state table of silicon and its effects on high-energy-density plasma simulations

Using density-functional theory–based molecular-dynamics simulations, we have investigated the equation of state for silicon in a wide range of plasma density and temperature conditions of ρ=0.001–500g/cm 3 and T=2000–10 8 K. With these calculations, we have established a first-principles equation-of-state (FPEOS) table of silicon for high-energy-density (HED) plasma simulations. When compared with the widely used SESAME-EOS model (Table 3810), we find that the FPEOS-predicted Hugoniot is ~20% softer; for off-Hugoniot plasma conditions, the pressure and internal energy in FPEOS are lower than those of SESAME EOS for temperatures above T ≈ 1–10 eV (depending on density), while the former becomes higher in the low- T regime. The pressure difference between FPEOS and SESAME 3810 can reach to ~50%, especially in the warm-dense-matter regime. Implementing the FPEOS table of silicon into our hydrocodes, we have studied its effects on Si-target implosions. When compared with the one-dimensional radiation-hydrodynamics simulation using the SESAME 3810 EOS model, the FPEOS simulation showed that (1) the shock speed in silicon is ~10% slower; (2) the peak density of an in-flight Si shell during implosion is ~20% higher than the SESAME 3810 simulation; (3) the maximum density reached in the FPEOS simulation is ~40% higher at the peak compression; and (4) the final areal density and neutron yield are, respectively, ~30% and ~70% higher predicted by FPEOS versus the traditional simulation using SESAME 3810. All of these features can be attributed to the larger compressibility of silicon predicted by FPEOS. Furthermore, these results indicate that an accurate EOS table, like the FPEOS presented here, could be essential for the precise design of targets for HED experiments.

36 MATERIALS SCIENCE↗
DOE OSTI2018

Simulation of Heavy-Duty Vehicles in Platooning Scenarios

Connectivity and automation provide the potential to use information about the environment and future driving to minimize energy consumption. In order to achieve this goal, the designers of control strategies need to simulate a wide range of driving situations that can interact with other vehicles and the infrastructure to account for the specific powertrain characteristics of each automated vehicle. We present here a simulation framework called RoadRunner, which aims to facilitate the design of powertrain-aware eco-driving algorithms and a better understanding of the interactions between automation and powertrain technology. RoadRunner allows users to simulate both powertrain and longitudinal dynamics within a simulated environment. The user defines the scenario to be simulated by providing a route profile, intersection control types, number of vehicles, vehicle class, powertrain technology, connectivity, and automation level. RoadRunner then builds a Simulink diagram of the scenario, including the information flows between vehicles, and between vehicles and the road. After the simulation, the user can analyze the driving, component operations, and detailed energy consumption rates for each simulated vehicle. We present a case study on heavy-duty vehicles platooning. Using RoadRunner and detailed road data from digital maps, we quantify the impact of gap setting on fuel consumption, for a real-world route.

Kim, Namdoo↗
DOE OSTI2019

Parallel-in-Time Simulation Algorithm for Power Electronics: MMC-HVdc System

The complexity in simulating power electronics like modular multilevel converters (MMCs) requires simulation algorithms to speed-up the process. Existing simulation algorithms exploit spatial parallelism to speed-up simulation. With rise in complexity of power electronics and presence of increased number of states within them, there are limits in the speed-up using spatial parallelism. In this paper, a temporal parallelism algorithm based on parallel-in-time methods is developed for simulation of power-electronics-systems. The temporal parallelism algorithm is based on computation of power-electronics-states on coarse and fine time-steps using different models. The models of power-electronics-systems used in coarse and fine time-steps are average-value and detailed models, respectively. The updates to states on coarse time-step are computed serially and are used to initialize the states on fine time-step. The updates on fine time-step are computed in parallel. A translation method is proposed in this paper to update the states on fine time-step from the simulations in the coarse time-step, and vice-versa. The proposed algorithm is applied to simulate MMCs and is validated with respect to a detailed reference MMC model. The proposed algorithm recorded a speed-up of up to 3.47× with 5 cores.

42 ENGINEERING↗
DOE OSTI2024

HumoNet: A Framework for Realistic Modeling and Simulation of Human Mobility Network

Understanding, analyzing, and predicting human mobility and dynamics are valuable to solving pressing problems, developing effective plans, and prescribing timely remedies. As a computational approach, realistic human mobility simulations allow us to understand, analyze, and predict complex systems, including human societies. Accurate simulations rely on (1) the model that captures interactions and behaviors of myriad entities in our society and (2) the mapping of model instances to real-world entities. Taking this into account, this paper introduces the Human Mobility Network simulation framework (HumoNet), an integrated patterns of life (POL) simulation framework that leverages real-world data layers including transportation networks, points of interest, populations, popularity, and human trajectories. HumoNet is a data informed model in which agents are equipped with activities, locomotion, and planning capabilities. To simulate realistic kinematic maneuvers of individuals in transportation networks, HumoNet harnesses a microscopic traffic simulator that provides interaction among vehicles and traffic objects. In this paper, we describe the framework, outline our methodologies, and discuss the data processing and challenges of each data layer. Through experiments, we demonstrate that our simulations capture key features of human mobility by comparing them to the literature and real data using standard measures of human mobility (i.e., the radius of gyration, number of locations visited, level of exploration) and metrics scoring (i.e., Jensen-Shannon divergence). We envision that the synthetic data produced by HumoNet will serve as a benchmark for analyzing epidemics, deploying EV charging networks, and validating AI/ML tasks such as location prediction.

Kim, Joon-Seok↗
DOE OSTI2022

Accelerating Flash-X Simulations with Asynchronous I/O

Most high-fidelity physics simulation codes, such as Flash-X, need to save intermediate results (checkpoint files) to restart or gain insights into the evolution of the simulation. These simulation codes save such intermediate files synchronously, where computation is stalled while the data is written to storage. Depending on the problem size and computational requirements, this file write time can be a substantial portion of the total simulation time. In order to hide the I/O latency of checkpointing, asynchronous I/O methods have been introduced. These methods use background threads for performing I/O while the main threads continue with the simulation. The usage of background threads can compete for resources on the node as well as with communication. In this paper, we evaluate the overheads and the overall benefit of asynchronous I/O in HDF5 to simulations. Results from real-world high-fidelity simulations on the Summit supercomputer show that I/O operation is overlapped with application communication or computation or both, effectively hiding some or all of the I/O latency. Our evaluation shows that while using asynchronous I/O adds overhead to the application, the I/O time reduction is more significant, resulting in overall up to 1.5X performance speedup.

Jain, Rajeev↗