Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Parallel processing”

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 1,045 records · Page 58
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Hexagonal Distributed Embedded Energy Converters (HexDEECs)

Distributed Embedded Energy Converter Technologies (DEEC-Tec) is a new domain for marine renewable energy research that utilizes a conglomeration of small distributed embedded energy converters (DEECs) that, in aggregate, form larger metamaterial frameworks. These resulting DEEC-Tec metamaterials can then, in turn, be used to construct flexible ocean wave energy converters called flexWECs. DEEC-Tec enables flexWECs: (i) to be inherently broad-banded ocean wave frequency energy converters and (ii) to have an inherent lack of highly loaded rigid bodies. The DEEC-Tec domain also benefits the marine renewable energy domain by inherently availing ways that marine energy can be harvested and converted that heretofore has not yet been considered possible: real-time execution of transforming topologies (e.g., actively changing a flexWEC's shape and form) and morphologies (e.g., actively changing a flexWEC's stiffness and damping throughout its entire structure). Presented, is one specific type of DEEC, a HexDEEC, that shows promise in aiding the adoption and further development of the DEEC-Tec domain - it is a small energy transducer being developed by the United States National Renewable Energy Laboratory. The HexDEEC is a small (characteristic length approximating a centimeter) energy transducer that converts the dynamic deformations of an elastomer into electricity through a charging-discharging cycle of a capacitor whose capacitance is varied by those elastic deformations. The HexDEEC is composed of a hyperelastic hexagonal housing (nominally silicon rubber) with six electrodes on its inner faces. The upper three electrodes share the same charge while the lower three electrodes oppose the upper electrode charges. Externally, the HexDEEC has two arms extending away from the middle vertices of the hexagon. Via principles governing the relationship between electrical capacitance and electrical potential (voltage and charge), electricity is generated when the HexDEEC's arms are dynamically pulled or released under tensile loading as doing so causes the distance between the upper and lower sets of electrodes to change - varying the energy converter's overall capacitance. Analytical and numerical modeling have already been used to estimate the electrical energy produced by a HexDEEC. The cursory models approximate the HexDEEC as a parallel plate variable capacitor - simplifying from six to two opposing plates with a constant dielectric volume between those two plates. To account for the elastic HexDEEC material properties, software such as SolidWorks and STAR-CCM+ have been used to generate hyperelastic models; notably, Mooney-Rivlin based models. Individual HexDEECs have been fabricated by drawing uncured liquid silicon rubber into molds via vacuum pressure. To simplify manufacturing, HexDEEC sub-components - e.g., electrodes, wires - can be placed within those molds such that they are directly embedded into the hexagonal housing during the curing process. Furthermore, DEEC-Tec metamaterials can be created by interweaving or sequentially layering multiple HexDEEC strands together. The HexDEEC based metamaterial could then generate electricity through its gross deformations. Ultimately, HexDEECs represent a specific type of energy transducer that can be leveraged, by the DEEC-Tec domain, to create metamaterials used to construct novel flexWECs.

DEEC-Tec↗
DOE OSTI2024

Regional-scale fault-to-structure earthquake simulations with the EQSIM framework: Workflow maturation and computational performance on GPU-accelerated exascale platforms

Continuous advancements in scientific and engineering understanding of earthquake phenomena, combined with the associated development of representative physics-based models, is providing a foundation for high-performance, fault-to-structure earthquake simulations. However, regional-scale applications of high-performance models have been challenged by the computational requirements at the resolutions required for engineering risk assessments. The EarthQuake SIMulation (EQSIM) framework, a software application development under the US Department of Energy (DOE) Exascale Computing Project, is focused on overcoming the existing computational barriers and enabling routine regional-scale simulations at resolutions relevant to a breadth of engineered systems. This multidisciplinary software development—drawing upon expertise in geophysics, engineering, applied math and computer science—is preparing the advanced computational workflow necessary to fully exploit the DOE’s exaflop computer platforms coming online in the 2023 to 2024 timeframe. Achievement of the computational performance required for high-resolution regional models containing upward of hundreds of billions to trillions of model grid points requires numerical efficiency in every phase of a regional simulation. This includes run time start-up and regional model generation, effective distribution of the computational workload across thousands of computer nodes, efficient coupling of regional geophysics and local engineering models, and application-tailored highly efficient transfer, storage, and interrogation of very large volumes of simulation data. This article summarizes the most recent advancements and refinements incorporated in the workflow design for the EQSIM integrated fault-to-structure framework, which are based on extensive numerical testing across multiple graphics processing unit (GPU)-accelerated platforms, and demonstrates the computational performance achieved on the world’s first exaflop computer platform through representative regional-scale earthquake simulations for the San Francisco Bay Area in California, USA.

58 GEOSCIENCES↗
DOE OSTI2024

Part-scale microstructure prediction for laser powder bed fusion Ti-6Al-4V using a hybrid mechanistic and machine learning model

Laser powder bed fusion (LPBF) Ti-6Al-4V is widely studied for use in structural applications in aerospace and medical industries, but mechanical anisotropy and microstructural inhomogeneity prohibits its wider adoption. Although successful microstructure prediction models have been developed, a remaining challenge is their limited integration across length/time scales and validation by experimental studies. Here, this work proposes a physics-augmented machine learning surrogate model to unite predictions of LPBF temperature, β phase morphology and texture, and α/α’ formation into a single framework that is calibrated and validated with experiments. First, a phase field (PF) model of the martensitic β→α’ transformation is developed and calibrated using data from in-situ synchrotron cyclic heating/cooling studies quantifying the variation of α phase fraction with time. In parallel, an established finite difference-Monte Carlo (FDMC) model predicts the part-scale temperature profile and β grain formation during solidification. A dataset is developed using LPBF cyclic temperature descriptors from the FDMC model as inputs and corresponding α/α’ phase fraction and width from the PF model as outputs. Five machine learning (ML) regression models are tested and optimized, having mean absolute error in testing ≤ 4 %, and the k-nearest neighbors (KNN) model is selected as the best performing. The KNN model is called at the nodal level during post-processing of the FDMC model to replace and downscale the response of the PF model. The combined agility and accuracy of the hybrid FDMC-ML model enables part-scale microstructure predictions that can be further used for property predictions to accelerate AM process optimization.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Building commissioning costs and savings across three decades and 1500 North American buildings

B.V. Building commissioning (Cx) is a process for assuring efficient building operations that can be applied to new construction and existing buildings, resulting in energy and non-energy benefits. Quantifying the benefits of commissioning is challenging, but a 2009 study of 643 commercial buildings provided a solid initial data set to which we added 839 additional buildings for a significantly expanded and updated meta-analysis representing 34.7 million square meters (373 million square feet) of floor area. Since 2009 the commissioning industry has continued to grow, driven by building codes, utility programs, and rising awareness of commissioning benefits. In parallel, building controls have become more sophisticated, and analytics software has emerged to assist with commissioning. We find that delivery mechanism and market segment are key determinants of outcomes, although significant and cost-effective savings are found across the spectrum. Median primary energy savings for Cx projects in existing buildings ranged from 5 percent for those conducted under utility programs, 9 percent for monitoring-based commissioning utility programs (i.e., augmented with submetering and diagnostics), and 14 percent for Cx projects outside of utility programs. Across all project types, median savings ranged from 3 percent for the lodging market segment to 16 percent for public order and safety facilities. Outcomes did not vary significantly by building size or by market segment. Energy savings are rarely estimated for new construction commissioning. We found that the median costs of Cx were lower for the 2018 sample than for the 2009 sample—$\$$2.85 per square meter ($\$$0.26 per square foot) for existing buildings (a 33 percent reduction) and $\$$8.78 per square meter ($\$$0.82 per square foot) for new construction (a reduction of almost 50 percent). The median simple payback time for existing buildings was 1.7 years, with a 25th–75th percentile range of 0.8–3.5 years. Overall, this article summarizes these and other key findings, and discusses how the 2018 data reflects shifts in commissioning practice and outcomes.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2024

Inverse chirality-induced spin selectivity effect in chiral assemblies of π -conjugated polymers

Coupling of spin and charge currents to structural chirality in non-magnetic materials, known as chirality-induced spin selectivity, is promising for application in spintronic devices at room temperature. Although the chirality-induced spin selectivity effect has been identified in various chiral materials, its Onsager reciprocal process, the inverse chirality-induced spin selectivity effect, remains unexplored. Here we report the observation of the inverse chirality-induced spin selectivity effect in chiral assemblies of π-conjugated polymers. Using spin-pumping techniques, the inverse chirality-induced spin selectivity effect enables quantification of the magnitude of the longitudinal spin-to-charge conversion driven by chirality-induced spin selectivity in different chiral polymers. By widely tuning conductivities and supramolecular chiral structures via a printing method, we found a very long spin relaxation time of up to several nanoseconds parallel to the chiral axis. Furthermore, our demonstration of the inverse chirality-induced spin selectivity effect suggests possibilities for elucidating the puzzling interplay between spin and chirality, and opens a route for spintronic applications using printable chiral assemblies.

36 MATERIALS SCIENCE↗
DOE OSTI2022

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this article, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this paper, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Nucleation and growth of blue phase liquid crystals on chemically-patterned surfaces: a surface anchoring assisted blue phase correlation length

In condensed matter, the correlation length is an essential parameter which describes the distance over which a material maintains its structural properties. In liquid crystals, for the orientational order parameter this characteristic length is of the order of nanometers, while in solid crystals it can extend to macroscopic length scales. Here, we report the measurement of the correlation length, or persistence length, of the crystallographic orientation of blue phases (BPs)—chiral liquid crystals with long-range 3D-crystalline structures and submicron-sized lattice-parameters. These materials exhibit phase transformations that have been identified as the liquid analog of crystal-crystal martensitic transformations. In this work, we use liquid epitaxial growth to achieve spontaneous BP-crystal nucleation and subsequent growth. Specifically, we design patterned substrates made of a binary array of regions with different liquid crystal anchoring, which facilitate a uniform nucleation and growth of BP-crystals with (100)-lattice orientation and a simple cubic symmetry. Furthermore, our results indicate that this simple cubic BP, the so-called blue phase II (BPII), forms first on the patterned surface, thereby propagating the growth of domains in directions parallel or perpendicular to the patterned regions. These results are used to understand the emergence of a surface anchoring assisted BPII-correlation length, taken as the distance over which the BP preserves its lattice orientation, as a function of time and pattern characteristics. We found that BPII single crystals can be achieved on patterned regions whose lateral dimensions are equal to or larger than 10 μm, consistent with our measurements of the BPII-correlation length. This newly acquired understanding of the role of surfaces on the formation of BPs is then used to design processes that permit formation of macroscopically large mono-domain, single-crystal BPs by relying on significantly reduced patterned areas (only part of the area is patterned), a feature that could benefit the applications of this intriguing class of materials.

correlation length↗
DOE OSTI2021

Dependence of Solar Wind Proton Temperature on the Polarization Properties of Alfvénic Fluctuations at Ion-kinetic Scales

In this work, we use fluctuating magnetic helicity to investigate the polarization properties of Alfvénic fluctuations at ion-kinetic scales in the solar wind as a function of β p , the ratio of proton thermal pressure to magnetic pressure, and θ vB , the angle between the proton flow and local mean magnetic field, B 0 . Using almost 15 yr of Wind observations, we separate the contributions to helicity from fluctuations with wavevectors, k, quasi-parallel and oblique to B 0 , finding that the helicity of Alfvénic fluctuations is consistent with predictions from linear Vlasov theory. This result suggests that the nonlinear turbulent fluctuations at these scales share at least some polarization properties with Alfvén waves. Furthermore, we also investigate the dependence of proton temperature in the β p –θ vB plane to probe for possible signatures of turbulent dissipation, finding that it correlates with θ vB . The proton temperature parallel to B 0 is higher in the parameter space where we measure the helicity of right-handed Alfvénic fluctuations, and the temperature perpendicular to B 0 is higher where we measure left-handed fluctuations. This finding is inconsistent with the general assumption that by sampling different θ vB in the solar wind we can analyze the dependence of the turbulence distribution on θ kB , the angle between k and B 0 . After ruling out both instrumental and expansion effects, we conclude that our results provide new evidence for the importance of local kinetic processes that depend on θ vB in determining proton temperature in the solar wind.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Nodeman: A Node Management Tool For Hpc Clusters

NodeMan is a command line tool to manage nodes in an HPC cluster. At it's core, it is an extensible framework composed of bash scripting and GNU parallel. HPC System Administrator will find it useful in that it encapsulates desired functions and allows them to be assembled in a way familiar to administrators - through pipes. In fact, NodeMan functions can work with common command line tools as long as they use stdin/stdout. System Administrators can construct moderately complex logic and filtering on a compact command line that would normally require a substantial shell script. In the spirit of clush and pdsh, it is able to run commands remotely on nodes. Additionally, NodeMan is more flexible. For example, it can interact with IPMI and naturally processes node lists for orchestrating different tools. The library of useful pre-built functions is growing. System administrators can easily create new functions and make it their own.

Serr, ScottM↗
DOE OSTI2022

Avoiding excess computation in asynchronous evolutionary algorithms

Abstract Asynchronous evolutionary algorithms are becoming increasingly popular as a means of making full use of many processors while solving computationally expensive search and optimization problems. These algorithms excel at keeping large clusters fully utilized, but may sometimes inefficiently sample an excess of fast‐evaluating solutions at the expense of higher‐quality, slow‐evaluating ones. We have previously introduced a steady‐state parent selection strategy, SWEET (“Selection whilE EvaluaTing”), that sometimes selects individuals that are still being evaluated and allows them to reproduce early. We perform a takeover‐time analysis that confirms that this strategy gives slow‐evaluating individuals that have higher fitnesses an increased ability to multiply in the population. We also find that SWEET appears effective at improving optimization performance on problems in which solution quality is positively correlated with evaluation time. We evaluate our approach on six simulated real‐valued optimization problems and three real‐world applications: an autonomous vehicle controller problem that involves tuning a spiking neural network and two adversarial EA problems. We further evaluate SWEET versus a basic asynchronous process in a simulated setting. We present evidence that SWEET outperforms basic asynchronous processes in a use‐case in which performance is positively correlated with evaluation time, and performs comparably (and often better) than basic asynchronous processes in several use‐cases where performance is negatively correlated with evaluation time. That said, in the cases where performance and evaluation time are negatively correlated the variance of outcomes for SWEET is notably high.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Enhanced accuracy through ensembling of randomly initialized auto-regressive models for dynamical systems

Computational mechanics simulations using traditional finite element methods (FEM) require prohibitively expensive computational resources for real-time engineering applications, design optimization, and digital twin implementations. While machine learning (ML) surrogate models offer significant computational speedups, autoregressive ML models for time-dependent mechanical systems suffer from error accumulation that compromises long-term prediction reliability - a critical concern for engineering applications where accuracy over extended time horizons is essential for safety and performance assessments. Here, we propose a deep ensemble framework specifically designed to address this challenge in computational mechanics applications, where multiple ML surrogate models with random weight initializations are trained in parallel and their predictions aggregated during inference. This approach leverages statistical diversity to maximize information gain from a fixed set of training data and to mitigate error propagation, while maintaining the computational efficiency that makes ML surrogates attractive for engineering practice. We validate the framework on three representative problems spanning critical areas of computational mechanics: stress field evolution in heterogeneous microstructures under complex loading (relevant to advanced materials design and composite analysis), planetary-scale shallow water dynamics (applicable to environmental and geotechnical engineering), and Gray-Scott reaction-diffusion systems (relevant to mass transport and chemical process engineering). Across all test cases, the ensemble approach demonstrates consistent error reduction of 15-33% compared to individual models. The codes for this work are available on GitHub (https://github.com/Graham-Brady-Research-Group/AutoregressiveEnsemble_SpatioTemporal_Evolution).

autoregressive prediction↗
DOE OSTI2020

AI-optimized detector design for the future Electron-Ion Collider: the dual-radiator RICH case

Advanced detector R&D requires performing computationally intensive and detailed simulations as part of the detector-design optimization process. Here, we propose a general approach to this process based on Bayesian optimization and machine learning that encodes detector requirements. As a case study, we focus on the design of the dual-radiator Ring Imaging Cherenkov (dRICH) detector under development as a potential component of the particle-identification system at the future Electron-Ion Collider (EIC). The EIC is a US-led frontier accelerator project for nuclear physics, which has been proposed to further explore the structure and interactions of nuclear matter at the scale of sea quarks and gluons. We show that the detector design obtained with our automated and highly parallelized framework outperforms the baseline dRICH design within the assumptions of the current model. Our technique can be applied to any detector R&D, provided that realistic simulations are available.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2021

Considerations regarding the Use of Computer Vision Machine Learning in Safety-Related or Risk-Significant Applications in Nuclear Power Plants

With the advancements made to date in the field of artificial intelligence (AI), significant potential exists to utilize AI capabilities for nuclear power plant (NPP) applications. AI can replicate human decision making and it is usually faster and more accurate than humans. For implementations that impact critical NPP applications (e.g., safety-related or non-safety systems that potentially affect overall plant risk), a deeper safety analysis of the AI methods is necessary. AI applied to NPP operations could resemble the use of digital I&C (DI&C) because such applications involve digital computer hardware and custom-designed software that input plant data, execute complex software algorithms, and output the results to a system or licensed human operator to potentially provoke an action. For AI methods to be compliant with current safety requirements for DI&C, AI compatibility must be evaluated, and AI-related gaps may exist that prevent the prompt deployment of AI in NPPs. This effort aims to evaluate how example AI technologies align with the DI&C safety framework, and discusses how they could be analyzed, modeled, tested, and validated in a manner similar to typical DI&C technologies. Because AI is a broad field that encompasses areas such as machine learning (ML), natural language processing, and computer vision, this research focused on a subset of methods categorized as the computer vision ML (CVML) methods. This report explores two CVML use cases, gauge reading and fire watch, considered relevant to the DI&C standards, as they could play a safety-critical role. For the gauge reading use case, a CVML-enabled technology that can read gauges at oblique angles is utilized. For the fire watch use case, a CVML-enabled technology is utilized that migrates fire watch from a manual (human) approach to automated fire detection. These use cases are mainly intended to give context to the CVML system discussion. This effort assumes the worst-case scenario, with the CVML system being used to replace a safety-related or risk-significant system, thus requiring evaluation. Evaluating CVML against most of the relevant safety requirements for DI&C yielded several CVML-specific considerations due to the uniqueness of its characteristics in comparison with typical DI&C systems. For example, CVML models often employ commonly used (open-source) datasets, and it is not always possible to determine the level of overlap among open-source datasets. Therefore, the independence of the developed CVML models when demonstrating diversity is questionable, therefore creating vulnerability to common cause failure (CCF). The design verification process is also impacted since the data overlap could result in overestimation of the software validation and verification (V&V) performance results. Section 2 of this report evaluates a list of the identified CVML-specific characteristics and discusses the resulting considerations and potential solutions in the context of each referenced requirement. A summation is provided in Section 3. This report is not to be used as a guideline. It was developed to identify and consider issues in the implementation of ML technologies used to augment activities that may have a bearing on plant operation. The report draws parallels to the use of DI&C technologies, for which many standards are available to guide their use in nuclear plant operation. It considers the technologies and some of the potential implications of their use in safety-related applications but is not intended to address regulatory or licensing related issues.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2022

Mechanisms for the Generation of Complex Fracture Networks: Observations From Slant Core, Analog Models, and Outcrop

We use observations of hydraulic fractures in core, outcrop attributes of natural hydraulic fractures, and analogue models, to address how hydraulic fracture networks evolve. A slant core from the Wolfcamp Formation—an unconventional shale hydrocarbon reservoir in the Permian Basin of West Texas—collected within 18 m and 30 m of two hydraulically stimulated horizontal wells, provided an opportunity to examine hydraulic fractures directly. In approximately 183 m of core, 309 calcite-sealed natural opening-mode fractures and 375 hydraulic fractures were identified. Many hydraulic fractures in the core show complex morphology, including twist-hackle segmentation, diversion, and bifurcation; these structures most commonly develop at lithological bed boundaries and mechanical heterogeneities such as natural fractures and concretions. An outcrop of bed-parallel pavements in the Cretaceous Boquillas Formation in West Texas contains opening-mode fractures that likely formed by natural hydraulic fracturing. Fracture traces provide evidence of twist-hackle segmentation, and are typically associated with bed boundaries and preexisting bed-parallel stylolites. A laboratory study of hydraulic fracturing of 33 synthetic blocks of gypsum and hydrostone revealed fracture steps, diversions, twist hackles, and multiple overlapping fractures together with information on fracture growth directions. These complexities in the fracture network were dominantly nucleated at inclusions used to simulate pre-existing fractures, and as a result of mechanical heterogeneity introduced by the wellbore and perforations. Collectively, our results show that complex fracture networks are produced in hydraulic fracturing of self-sourced reservoir strata. Mechanical stratigraphic boundaries and other heterogeneities are likely to enhance fracture network complexity through the processes of segmentation, diversion, and bifurcation. These processes create multiple fracture strands, resulting in an increased number of hydraulic fractures over those initiated, thereby increasing total fracture surface area. Our study provides insight into hydraulic fracture network propagation, and has applications for evaluation, completion, production, and fracture modeling of unconventional reservoirs.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

C-SAW: a framework for graph sampling and random walk on GPUs

Many applications require to learn, mine, analyze and visualize large-scale graphs. These graphs are often too large to be addressed efficiently using conventional graph processing technologies. Fortunately, recent research efforts find out graph sampling and random walk, which significantly reduce the size of original graphs, can benefit the tasks of learning, mining, analyzing and visualizing large graphs by capturing the desirable graph properties. This paper introduces C-SAW, the first framework that accelerates Sampling and Random Walk framework on GPUs. Particularly, C-SAW makes three contributions: First, our framework provides a generic API which allows users to implement a wide range of sampling and random walk algorithms with ease. Second, offloading this framework on GPU, we introduce warp-centric parallel selection, and two novel optimizations for collision migration. Third, towards supporting graphs that exceed the GPU memory capacity, we introduce efficient data transfer optimizations for out-of-memory and multi-GPU sampling, such as workload-aware scheduling and batched multi-instance sampling. Taken together, our framework constantly outperforms the state of the art projects in addition to the capability of supporting a wide range of sampling and random walk algorithms.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

HydraGNN_Predictive_GFM_2024 - Ensemble of predictive graph foundation models for ground state atomistic materials modeling

We provide the ensemble of fifteen pre-trained graph foundation models (GFMs) for atomistic materials modeling applications. Each one of the fifteen GFMs has been trained on five open-source datasets that (once aggregated) amount to over 154 million atomistic structures, which cover over two-thirds of the natural elements of the periodic table and that comprises a broad set of organic and inorganic compounds. This vast set of atomistic structures comprises ground state configurations that are dynamically stable (i.e., equilibrated structures with atomic forces approximately close to zero values) as well as dynamically unstable structures (i.e., non-equilibrium structures with non-negligible non-zero values of atomic forces). The ensemble of datasets aggregated does NOT include excited states. The datasets have been curated to remove atomistic structures with spectral norm of the force tensor above 100 eV/angstrom. Moreover, a linear term of the energy was computed for each dataset using a linear regression model that uses the chemical concentration of each natural element as regressor. The linear term predicted by the linear regression model has been subtracted from each original energy value to perform a re-alignment of the energy values across different electronic structures approximation theories performed to generate the diverse multi-source, multi-fidelity datasets. The folder "ADIOS_files" contains the set of pre-processed datasets in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used for the development and training of GFMs in this work. The "ADIOS_files" directory contains 6 sub-directories named as follows: - ANI1x-v3.bp - MPTrj-v3.bp - OC2020-20M-v3.bp - OC2020-v3.bp - OC2022-v3.bp - qm7x-v3.bp Each sub-directory contains the pre-processed datasets converted in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used to the development, training, and performance testing of the ensemble go predictive graph foundation models. Each GFM was developed using HydraGNN (https://github.com/ORNL/HydraGNN) as underlying graph neural network (GNN) architecture. The multi-task learning (MTL) capability of HydraGNN was used to simultaneously train the GFMs on labeled values for direct predictions of energy (a total system property of an atomistic structure that measures the chemical stability) and atomic forces (an atomic level property of an atomistic structure that measures the dynamical stability). The hyper parameters of the GFM have been tuned using scalable hyperparameter optimization (HPO) algorithms implemented in the software DeepHyper (https://github.com/deephyper/deephyper). The pre-training of each HPO trial was performed using distributed data parallelism (DDP) to scale the training across 128 compute nodes of the exascale OLCF supercomputer Frontier. Each HPO trial was trained only for 10 epochs and an early stopping was performed to avoid wasting significant computational resources on GNN architectures that were clearly underperforming. For each HPO trial, the 'omnistat' tool developed by (AMD Research - Advanced Micro Device) was used to measure the total energy consumption in kWh. The ensemble of GFMs was obtained by selecting the fifteen best performing HPO trials. Four models have been selected for their clear advantage in accuracy, and these are the GFMs with IDs 229, 156, 147, 260. Additional eleven models have been selected based on judicious balance between accuracy and energy consumption needed for training, and these are the GFMs with IDs 165, 78, 137, 1, 175, 171, 181, 67, 179, 167, 351. Each selected GFM of the ensemble was continued to cumulate a total of at most 30 epochs. In some cases, the total number of epochs actually performed was les than 30 due to two combined factors: (1) the size of the GFM (i.e., the number of model parameters to train) and (2) the total wall-clock time for which the computational resources could be allocated on OLCF-Frontier. The "Ensemble_of_models" directory contains 15 sub-directories named as follows: - gfm_0.229 - gfm_0.156 - gfm_0.147 - gfm_0.260 - gfm_0.165 - gfm_0.78 - gfm_0.137 - gfm_0.1 - gfm_0.175 - gfm_0.171 - gfm_0.181 - gfm_0.67 - gfm_0.179 - gfm_0.167 - gfm_0.351 Each one of these sub-directories refers to one of the fifteen HPO trials that have been selected to continue the pre-training with at most 30 epochs. With each sub-directory associated with a specific HPO trial, the following files can be found: - config.json: file for argument parsing to develop and train an HydraGNN architecture - gfm_0.ID_epoch_N.pk: file with model parameters for HPO ID trial after N epochs of training The ensemble of fifteen GFM architectures was used for (1) ensemble averaging to stabilize the predictions of energy and atomic forces after pre-training for post-processing analysis and (2) ensemble uncertainty quantification (UQ). The code used to develop, pre-train, and load the pre-trained models for post-processing analysis is available on the ORNL-GitHub at the following link: https://github.com/ORNL/HydraGNN/tree/Predictive_GFM_2024

36 MATERIALS SCIENCE↗
DOE OSTI2022

Neuromorphic scaling advantages for energy-efficient random walk computations

Neuromorphic computing, which aims to replicate the computational structure and architecture of the brain in synthetic hardware, has typically focused on artificial intelligence applications. What is less explored is whether such brain-inspired hardware can provide value beyond cognitive tasks. Here we show that the high degree of parallelism and configurability of spiking neuromorphic architectures makes them well suited to implement random walks via discrete-time Markov chains. Overall, these random walks are useful in Monte Carlo methods, which represent a fundamental computational tool for solving a wide range of numerical computing tasks. Using IBM’s TrueNorth and Intel’s Loihi neuromorphic computing platforms, we show that our neuromorphic computing algorithm for generating random walk approximations of diffusion offers advantages in energy-efficient computation compared with conventional approaches. We also show that our neuromorphic computing algorithm can be extended to more sophisticated jump-diffusion processes that are useful in a range of applications, including financial economics, particle physics and machine learning.

97 MATHEMATICS AND COMPUTING↗