Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “MPI characterization”

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.

49 records · Page 3

DAMAGE MODELING OF INJECTION-MOLDED SHORT- AND LONG-FIBER THERMOPLASTICS

This article applies the recent anisotropic rotary diffusion – reduced strain closure (ARD-RSC) model for predicting fiber orientation and a new damage model for injection-molded long-fiber thermoplastics (LFTs) to analyze progressive damage leading to total failure of injection-molded long-glass-fiber/polypropylene (PP) specimens. The ARD-RSC model was implemented in a research version of the Autodesk Moldflow Plastics Insight (MPI) processing code, and it has been used to simulate injection-molding of a long-glass-fiber/PP plaque. The damage model combines micromechanical modeling with a continuum damage mechanics description to predict the nonlinear behavior due to plasticity coupled with damage in LFTs. This model has been implemented in the ABAQUS finite element code via user-subroutines and has been used in the damage analyses of tensile specimens removed from the injection-molded long-glass-fiber/PP plaques. Experimental characterization and mechanical testing were performed to provide input data to support and validate both process modeling and damage analyses. The predictions are in agreement with the experimental results.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

ScalaTrace: Tracing, Analysis and Modeling of HPC Codes at Scale

Characterizing the communication behavior of large-scale applications is a difficult and costly task due to code/system complexity and their long execution times. An alternative to running actual codes is to gather their communication traces and then replay them, which facilitates application tuning and future procurements. While past approaches lacked lossless scalable trace collection, we contribute an approach that provides orders of magnitude smaller, if not near constant-size, communication traces regardless of the number of nodes while preserving structural information. We introduce intra- and inter-node compression techniques of MPI events, we develop a scheme to preserve time and causality of communication events, and we present results of our implementation for BlueGene/L. Given this novel capability, we discuss its impact on communication tuning and on trace extrapolation. To the best of our knowledge, such a concise representation of MPI traces in a scalable manner combined with time-preserving deterministic MPI call replay are without any precedence.

97 MATHEMATICS AND COMPUTING↗

Characterization and identification of HPC applications at leadership computing facility

High Performance Computing (HPC) is an important method for scientific discovery via large-scale simulation, data analysis, or artificial intelligence. Leadership-class supercomputers are expensive, but essential to run large HPC applications. The Petascale era of supercomputers began in 2008, with the first machines achieving performance in excess of one petaflops, and with the advent of new supercomputers in 2021 (e.g., Aurora, Frontier), the Exascale era will soon begin. However, the high theoretical computing capability (i.e., peak FLOPS) of a machine is not the only meaningful target when designing a supercomputer, as the resources demand of applications varies. A deep understanding of the characterization of applications that run on a leadership supercomputer is one of the most important ways for planning its design, development and operation. In order to improve our understanding of HPC applications, user demands and resource usage characteristics, we perform correlative analysis of various logs for different subsystems of a leadership supercomputer. This analysis reveals surprising, sometimes counter-intuitive patterns, which, in some cases, conflicts with existing assumptions, and have important implications for future system designs as well as supercomputer operations. For example, our analysis shows that while the applications spend significant time on MPI, most applications spend very little time on file I/O. Combined analysis of hardware event logs and task failure logs show that the probability of a hardware FATAL event causing task failure is low. Combined analysis of control system logs and file I/O logs reveals that pure POSIX I/O is used more widely than higher level parallel I/O. Based on holistic insights of the application gained through combined and co-analysis of multiple logs from different perspectives and general intuition, we engineer features to "fingerprint" HPC applications. We use t-SNE (a machine learning technique for dimensionality reduction) to validate the explainability of our features and finally train machine learning models to identify HPC applications or group those with similar characteristic. To the best of our knowledge, this is the first work that combines logs on file I/O, computing, and inter-node communication for insightful analysis of HPC applications in production.

Liu, Zhengchun↗

Uncovering grain and subgrain microstructure at the scale of additive manufacturing melt tracks with a scalable cellular automaton solidification model

Metal additive manufacturing, characterized by rapid solidification, yields refined grains with a distinctive cellular subgrain microstructure that plays a pivotal role in determining material properties. Due to the significant computational expense demanded to simulate the required physics with submicron spatial resolution, their numerical simulations have been limited to proof-of-concept studies to either 2D or small subregions of a melt pool. In this study, an open-source, scalable, solidification code, muMatScale, based on the cellular automaton method, has been developed to predict the grain and the underlying subgrain microstructure over an entire melt pool. The model incorporates flexible parallelization schemes, utilizing MPI and OpenMP GPU Offloading, in addition to appropriate multi-physics specific to non-equilibrium rapid solidification in AM. The impact of nucleation parameters on grain microstructures was investigated with a focus on grain size variations and morphology transitions. With selected nucleation parameters, the simulation predicted the grain size, subgrain morphology, crystallographic orientation, and microsegregation aligned with experimental measurements. The model demonstrates that epitaxial grain growth is a dominant factor at the melt pool boundary, influencing grain size variation under different grain sizes in the build plate while maintaining consistent primary dendrite arm spacing under identical thermal conditions. Here, the highly efficient numerical model enables large-scale simulations with a spatial resolution of 100 nm or less, unveiling unprecedented insights into thermal and solutal diffusion driven grain growth, and the subgrains with microsegregation within grains in 3D across scales. muMatScale will enable the linking of submicron length-scale microstructure to part-level material behavior by investigating fundamental solidification problems at the intercellular scale in many-track and many-layer builds.

36 MATERIALS SCIENCE↗

ERF: Energy Research and Forecasting

The Energy Research and Forecasting (ERF) code is a new model that simulates the mesoscale and microscale dynamics of the atmosphere using the latest high-performance computing architectures. It employs hierarchical parallelism using an MPI+X model, where X may be OpenMP on multicore CPU-only systems, or CUDA, HIP, or SYCL on GPU-accelerated systems. ERF is built on AMReX (Zhang et al., 2019, 2021), a block-structured adaptive mesh refinement (AMR) software framework that provides the underlying performance-portable software infrastructure for block-structured mesh operations. The "energy" aspect of ERF indicates that the software has been developed with renewable energy applications in mind. In addition to being a numerical weather prediction model, ERF is designed to provide a flexible computational framework for the exploration and investigation of different physics parameterizations and numerical strategies, and to characterize the flow field that impacts the ability of wind turbines to extract wind energy. The ERF development is part of a broader effort led by the US Department of Energy's Wind Energy Technologies Office.

17 WIND ENERGY↗

Three practical workflow schedulers for easy maximum parallelism

Runtime scheduling and workflow systems are an increasingly popular algorithmic component in HPC because they allow full system utilization with relaxed synchronization requirements. There are so many special-purpose tools for task scheduling, one might wonder why more are needed. Use cases seen on the Summit supercomputer needed better integration with MPI and greater flexibility in job launch configurations. Preparation, execution, and analysis of computational chemistry simulations at the scale of tens of thousands of processors revealed three distinct workflow patterns. A separate job scheduler was implemented for each one using extremely simple and robust designs: file-based, task-list based, and bulk-synchronous. Comparing to existing methods shows unique benefits of this work, including simplicity of design, suitability for HPC centers, short startup time, and well-understood per-task overhead. All three new tools have been shown to scale to full utilization of Summit, and have been made publicly available with tests and documentation. This work presents a complete characterization of the minimum effective task granularity for efficient scheduler usage scenarios. Here, these schedulers have the same bottlenecks, and hence similar task granularities as those reported for existing tools following comparable paradigms.

97 MATHEMATICS AND COMPUTING↗

Oxyanion Induced Variations in Domain Structure for Amorphous Cobalt Oxide Oxygen Evolving Catalysts, Resolved by X-ray Pair Distribution Function Analysis

Amorphous thin film oxygen evolving catalysts, OECs, of first-row transition metals show promise to serve as self-assembling photoanode materials in solar-driven, photoelectrochemical `artificial leaf' devices. This report demonstrates the ability to use high-energy X-ray scattering and atomic pair distribution function analysis, PDF, to resolve structure in amorphous metal oxide catalyst films. The analysis is applied here to resolve domain structure differences induced by oxyanion substitution during the electrochemical assembly of amorphous cobalt oxide catalyst films, Co-OEC. PDF patterns for Co-OEC films formed using phosphate, Pi, methylphosphate, MPi, and borate, Bi, electrolyte buffers show that the resulting domains vary in size following the sequence Pi < MPi < Bi. The increases in domain size for CoMPi and CoBi were found to be correlated with increases in the contributions from bilayer and trilayer stacked domains having structures intermediate between those of the LiCoOO and CoO(OH) mineral forms. The lattice structures and offset stacking of adjacent layers in the partially stacked CoMPi and CoBi domains were best matched to those in the LiCoOO layered structure. The results demonstrate the ability of PDF analysis to elucidate features of domain size, structure, defect content and mesoscale organization for amorphous metal oxide catalysts that are not readily accessed by other X-ray techniques. Finally, PDF structure analysis is shown to provide a way to characterize domain structures in different forms of amorphous oxide catalysts, and hence provide an opportunity to investigate correlations between domain structure and catalytic activity.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Direct Numerical Simulation of Autoignition in a Jet in a Cross-Flow Configuration: ALCF-2 Early Science Program (Technical Report)

Autoignition in turbulent flows is a challenging fundamental problem due to the intricate coupling of different physical and chemical processes extending over multiple flow and chemistry scales. At the same time, the improved understanding and ability to predict autoignition in flows characterized by considerable fluctuations of velocity, composition, and temperature is essential for the development of novel low-emission concepts for power generation. The aim of this project is to study the fundamental aspects of autoignition in a fuel-air mixing device directly applicable to mixing ducts in gas turbines. The NEK5000-based code for low Mach number reactive flows is used to perform very large scale direct numerical simulations of autoignition of a diluted hydrogen jet ejected in a cross-flowing stream of hot turbulent air in a laboratory-scale configuration. We report on our experience running NEK5000 on the new BGQ system at ALCF mira during the early science period (ESP). First of all, the most efficient problem size per MPI-rank is obtained through core-level efficiency metric measured from the target simulation. Furthermore, the most efficient number of ranks is found through strong scaling experiments. Finally, low-level insight into the observed parallel efficiency is enabled through IBM’s HPC Toolkit libraries.

Abdilghanie, A.↗

Intracellular Dynamics of Superparamagnetic Iron Oxide Nanoparticles for Magnetic Partical Imaging

Superparamagnetic iron oxide nanoparticles (SPIONs) are a foundational platform for a variety of biomedical applications. Of particular interest is Magnetic Particle Imaging (MPI), which is a growing area of research and development due to its advantages including high resolution and sensitivity with positive contrast. There has been signicant work in the area of in vivo optimization of SPIONs for MPI as well as their biodistribution in and clearance from the body. However, little is known about the dynamics of SPIONs following cellular internalization which may limit their usefulness in a variety of potential imaging and treatment applications. This work shows a clear 20% decrease in magnetic performance of SPIONs, as observed by Magnetic Particle Spectroscopy (MPS), after internalization and systematic consideration of applicable factors that aect SPION signal generation, including microstructure, environment, and interparticle interactions. There is no observed change to SPION microstructure after internalization, and the surrounding environment plays little to no role in magnetic response for the SPIONs studied here. Interparticle interactions described by dipole-dipole coupling of SPIONs held close to one another after internalization are shown to be the dominant cause of decreased magnetic performance in cells. These conclusions were drawn from transmission electron microscopy (TEM) image analysis at relevant length scales, experimentally prepared and characterized SPIONs in varied environmental conditions, and theoretical modeling with Monte Carlo simulations.

Teeman, Eric M.↗

Monte Carlo modeling and analyses of YALINA- booster subcritical assembly Part II : pulsed neutron source.

One of the most reliable experimental methods for measuring the kinetic parameters of a subcritical assembly is the Sjoestrand method applied to the reaction rate generated from a pulsed neutron source. This study developed a new analytical methodology for characterizing the kinetic parameters of a subcritical assembly using the Sjoestrand method, which allows comparing the analytical and experimental time dependent reaction rates and the reactivity measurements. In this methodology, the reaction rate, detector response, is calculated due to a single neutron pulse using MCNP/MCNPX computer code or any other neutron transport code that explicitly simulates the fission delayed neutrons. The calculation simulates a single neutron pulse over a long time period until the delayed neutron contribution to the reaction is vanished. The obtained reaction rate is superimposed to itself, with respect to the time, to simulate the repeated pulse operation until the asymptotic level of the reaction rate, set by the delayed neutrons, is achieved. The superimposition of the pulse to itself was calculated by a simple C computer program. A parallel version of the C program is used due to the large amount of data being processed, e.g. by the Message Passing Interface (MPI). The new calculation methodology has shown an excellent agreement with the experimental results available from the YALINA-Booster facility of Belarus. The facility has been driven by a Deuterium-Deuterium or Deuterium-Tritium pulsed neutron source and the (n,p) reaction rate has been experimentally measured by a {sup 3}He detector. The MCNP calculation has utilized the weight window and delayed neutron biasing variance reduction techniques since the detector volume is small compared to the assembly volume. Finally, this methodology was used to calculate the IAEA benchmark of the YALINA-Booster experiment.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Transient Triplet Metallopnictinidenes M–Pn (M = Pd II , Pt II ; Pn = P, As, Sb): Characterization and Dimerization

Nitrenes (R–N) have been subject to a large body of experimental and theoretical studies. The fundamental reactivity of this important class of transient intermediates has been attributed to their electronic structures, particularly the accessibility of triplet vs singlet states. In contrast, electronic structure trends along the heavier pnictinidene analogues (R–Pn; Pn = P–Bi) are much less systematically explored. We here report the synthesis of a series of metallodipnictenes, {M–Pn=Pn–M} (M = Pd II , Pt II ; Pn = P, As, Sb, Bi) and the characterization of the transient metallopnictinidene intermediates, {M–Pn} for Pn = P, As, Sb. Structural, spectroscopic, and computational analysis revealed spin triplet ground states for the metallopnictinidenes with characteristic electronic structure trends along the series. In comparison to the nitrene, the heavier pnictinidenes exhibit lower-lying ground state SOMOs and singlet excited states, thus suggesting increased electrophilic reactivity. Furthermore, the splitting of the triplet magnetic microstates is beyond the phosphinidenes {M–P} dominated by heavy pnictogen atom induced spin–orbit coupling.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Pilot Implementation of a Field Study Design to Evaluate the Impact of Source Control Measures on Indoor Air Quality in High Performance Homes

To improve the indoor air quality in new, high performance homes, a variety of standards and rating programs have been introduced to identify building materials that are designed to have lower emission rates of key contaminants of concern and a number of building materials are being introduced that are certified to these standards. For example, the U.S. Department of Energy (DOE) Zero Energy Ready Home program requires certification under the U.S. Environmental Protection Agency (EPA) Indoor airPLUS (IaP) label, which requires the use of PS1 or PS2 certified plywood and OSB; low-formaldehyde emitting wood products; low- or no-VOC paints and coatings as certified by Green Seal Standard GS-11, GreenGuard, SCS Indoor Advantage Gold Standard, MPI Green Performance Standard, or another third party rating program; and Green Label-certified carpet and carpet cushions. However, little is known regarding the efficacy of the IAP requirements in measurably reducing contaminant exposures in homes. The goal of this project is to develop a robust experimental approach and collect preliminary data to support the evaluation of indoor air quality (IAQ) measures linked to IAP-approved low-emitting materials and finishes in new residential homes. To this end, the research team of Pacific Northwest National Laboratory (PNNL) and Lawrence Berkeley National Laboratory (LBNL) developed a detailed experimental plan to measure IAQ constituents and other parameters, over time, in new homes constructed with materials compliant with IAP’s low-emitting material and ventilation requirements (i.e., section 6.1, 6.2, 6.3, and 7.2) and similar homes constructed to the state building code with conventional materials. The IAQ in IAP and conventional homes of similar age, location, and construction style is quantified as the differences in the speciated VOC and aldehyde concentrations, normalized to dilution rates. The experimental plan consists of methods to evaluate the difference between low-emitting and “conventional” materials as installed in newly constructed residential homes using both (1) highly controlled, short-term active samples to precisely characterize the building-related chemical emissions and building contents and (2) a week-long passive sample designed to capture the impact of occupant behavior and related activities on measured IAQ contaminant levels indoors. The combination of detailed short-term measurements with the home under controlled/consistent conditions during pre- and post-occupancy and the week-long passive sampling data provide the opportunity to begin to separate the different emission sources and help isolate and quantify variability in the monitored homes. Between April and August 2014, the research team performed pre-occupancy and post-occupancy sampling in one conventional home and two homes built with low-emitting materials that were generally consistent with EPA’s Indoor airPLUS guidelines. However, for a number of reasons, the full experimental plan was not implemented. The project was intended to continue for up to three years to asses long-term changes in IAQ but the project was limited to one calendar year. As a result, several of the primary research questions related to seasonal impacts and the long-term trends in IAQ could not be addressed. In addition, there were several unexpected issues related to recruiting, availability of home types, and difficulty coordinating with builders/realtors/homeowners. Several field monitoring issues also came up that provide “lessons learned” that led to improvements to the original monitoring plan. The project produced a good experimental plan that is expected to be be useful for future efforts collecting data to support answering these same or similar research questions.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Effects of pulse duration on magnetostimulation thresholds

Purpose: Medical imaging techniques such as magnetic resonance imaging and magnetic particle imaging (MPI) utilize time-varying magnetic fields that are subject to magnetostimulation limits, which often limit the speed of the imaging process. Various human-subject experiments have studied the amplitude and frequency dependence of these thresholds for gradient or homogeneous magnetic fields. Another contributing factor was shown to be number of cycles in a magnetic pulse, where the thresholds decreased with longer pulses. The latter result was demonstrated on two subjects only, at a single frequency of 1.27 kHz. Hence, whether the observed effect was due to the number of cycles or due to the pulse duration was not specified. In addition, a gradient-type field was utilized; hence, whether the same phenomenon applies to homogeneous magnetic fields remained unknown. Here, the authors investigate the pulse duration dependence of magnetostimulation limits for a 20-fold range of frequencies using homogeneous magnetic fields, such as the ones used for the drive field in MPI. Methods: Magnetostimulation thresholds were measured in the arms of six healthy subjects (age: 27 ± 5 yr). Each experiment comprised testing the thresholds at eight different pulse durations between 2 and 125 ms at a single frequency, which took approximately 30–40 min/subject. A total of 34 experiments were performed at three different frequencies: 1.2, 5.7, and 25.5 kHz. A solenoid coil providing homogeneous magnetic field was used to induce stimulation, and the field amplitude was measured in real time. A pre-emphasis based pulse shaping method was employed to accurately control the pulse durations. Subjects reported stimulation via a mouse click whenever they felt a twitching/tingling sensation. A sigmoid function was fitted to the subject responses to find the threshold at a specific frequency and duration, and the whole procedure was repeated at all relevant frequencies and pulse durations. Results: The magnetostimulation limits decreased with increasing pulse duration (T{sub pulse}). For T{sub pulse} < 18 ms, the thresholds were significantly higher than at the longest pulse durations (p < 0.01, paired Wilcoxon signed-rank test). The normalized magnetostimulation threshold (B{sub Norm}) vs duration curve at all three frequencies agreed almost identically, indicating that the observed effect is independent of the operating frequency. At the shortest pulse duration (T{sub pulse} ≈ 2 ms), the thresholds were approximately 24% higher than at the asymptotes. The thresholds decreased to within 4% of their asymptotic values for T{sub pulse} > 20 ms. These trends were well characterized (R{sup 2} = 0.78) by a stretched exponential function given by B{sub Norm}=1+αe{sup −(T{sub p}{sub u}{sub l}{sub s}{sub e}/β){sup γ}}, where the fitted parameters were α = 0.44, β = 4.32, and γ = 0.60. Conclusions: This work shows for the first time that the magnetostimulation thresholds decrease with increasing pulse duration, and that this effect is independent of the operating frequency. Normalized threshold vs duration trends are almost identical for a 20-fold range of frequencies: the thresholds are significantly higher at short pulse durations and settle to within 4% of their asymptotic values for durations longer than 20 ms. These results emphasize the importance of matching the human-subject experiments to the imaging conditions of a particular setup. Knowing the dependence of the safety limits to all contributing factors is critical for increasing the time-efficiency of imaging systems that utilize time-varying magnetic fields.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗