SEARCH · Engineering Papers

Results for “complex computing”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 937 records · Page 52

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Exploring Architectural-Aware Affinity Policies in Modern HPC Runtimes

Modern commodity and High-Performance Computing (HPC) systems are evolving with complex CPU architectures. These architectures now feature higher core and NUMA domain counts and implement features such as hyperthreading. When considering significant differences in hardware configurations, library availability, and hardware-tailored system/software stacks, which could substantially vary from one system to another, performance portability is hard to achieve. Throughout the years, this trend resulted in an increasingly high burden on application developers to fine-tune their workloads for each architecture. This work explores how hardware-dependent aspects such as locality/process/thread affinity affect performance in modern CPU architectures. We focus our study on the Global Memory and Threading (GMT) distributed runtime system as a representative of Partitioned Global Address Space (PGAS) software stacks commonly adopted for productivity. In particular, to appreciate performance implications, we evaluate GMT’s thread affinity policies, and, introduce two new ones which exploit architectural awareness. Finally, we explore alternative NUMA configurations via different process bindings and perform a scalability study on three HPC clusters with varying CPU architectures and NUMA layouts. Our analysis indicates that more complex architectures are more affected by affinity and binding policies and highlights the importance of setting proper runtime configurations to achieve superior performance.

Di Dio Lavore, Ian↗

DOE OSTI2024

SmartFuse: Reconfigurable Smart Switches to Accelerate Fused Collectives in HPC Applications

Communication switches have sometimes been augmented to process collectives (e.g., the IBM BlueGene project and the Mellanox SHArP switch). In this work, we find that there is a great acceleration opportunity through the further augmentation of switches to accelerate more complex functions that combine communication with computation. We consider three types of such functions. The first is fully-fused collectives built by fusing multiple existing collectives like Allreduce with Alltoall. The second is semi-fused collectives built by combining a collective with another computation. The third we refer to as higher-order collectives built by combining multiple computations and communications, such as to perform matrix-matrix multiply (PGEMM). In this work, we propose a framework called SmartFuse to accelerate fused collective functions. The core of SmartFuse is a reconfigurable smart switch to support these operations. The semi/fully fused collectives are implemented with a CGRAlike architecture, while higher-order collectives are implemented with a more specialized computational unit that can also schedule communication. Supporting our framework is software to evaluate and translate relevant parts of the input program, compile them into a control data flow graph, and then map this graph to the switch hardware. The proposed framework, once deployed, has the strong potential to accelerate existing HPC applications transparently by encapsulation within an MPI implementation. Experimental results show that this approach improves the performance of the PGEMM kernel, MINIFE, and AMG by, on average, 94%, 15%, and 13%, respectively.

Haghi, Pouya↗

DOE OSTI2023

Improving Mathematical Exposition of an Industrial-Scale Linear Program

Industrial-scale models require considerable setup time; hence, once built, they are used in myriad ways to consider closely related cases. In practice, the code for these models frequently evolves without appropriate notational choices, largely as a result of the lengthy development time of, and the number of individuals contributing to, their formulation. This leads to inefficiencies and obfuscates model structures that might be leveraged to expedite solutions. In this paper, we advocate for an emerging literature on model formulation “best practices” and present the reformulation of a widely used industrial-scale linear program. The efficient mathematical expression of this linear program, used to plan capacity expansion in the energy sector, allows for greater transparency of model structures and enhanced ability to identify computational performance improvements, as well as a lucid interpretation of its solutions. This type of formulation is employed in several mathematical programming courses at our university as an example of the advantages of best practices; the model more broadly is used widely to inform policy in the U.S. energy sector.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Simultaneous imaging of two-phase velocities in particle-laden flows by two-color optical phase discrimination

In this Letter, we present a particle image/tracking velocimetry (PIV/PTV) technique for simultaneous velocity measurement of both fluid and particle phases, adopting newly developed optical phase discrimination methods and novel optical particles. Spherical acrylic (PMMA) particles of diameter ∼ <#comment/> O ( 100 µ <#comment/> m ) were used as the particle phase, while fine B A M : E u 2 + phosphors of diameter ∼ <#comment/> O ( 1 µ <#comment/> m ) were used as the fluid tracer. Under Nd:YAG 355 nm laser excitation, both the laser-induced fluorescence (LIF) from PMMA and laser-induced phosphorescence (LIP) from B A M : E u 2 + provided sufficiently strong signals for PIV imaging with two non-intensified cameras and were clearly separable for phase discrimination using spectral filters and temporal profiles. The advantages of the PIV/PTV method include the relatively low cost of B A M : E u 2 + phosphors, high sphericity and narrow size distribution of PMMA particles with LIF emission, and direct optical discrimination eliminating artifacts, while requiring much less computational capacity for PIV/PTV processing of complex particle-laden flows.

Bi, Xiaopeng (ORCID:0000000190847325)↗

DOE OSTI2025

Metalens formed by structured arrays of atomic emitters

Abstract Arrays of atomic emitters have proven to be a promising platform to manipulate and engineer optical properties, due to their efficient cooperative response to near‐resonant light. Here, we theoretically investigate their use as an efficient metalens. We show that, by spatially tailoring the (subwavelength) lattice constants of three consecutive two‐dimensional arrays of identical atomic emitters, one can realize a large transmission coefficient with arbitrary position‐dependent phase shift, whose robustness against losses is enhanced by the collective response. To characterize the efficiency of this atomic metalens, we perform large‐scale numerical simulations involving a substantial number of atoms (N∼ 5 × 10 5 ) that is considerably larger than comparable works. Our results suggest that low‐loss, robust optical devices with complex functionalities, ranging from metasurfaces to computer‐generated holograms, could be potentially assembled from properly engineered arrays of atomic emitters.

Materials Science↗

DOE OSTI2021

Assimilation of Multiscale Data into Multifidelity Biogeochemical Models (Final Report)

Quantitative predictions of subsurface processes rely on computational models that capture, with different degrees of fidelity, complex interactions between hydrologic and biogeochemical processes. Molecular- and pore-scale models provide a high-fidelity representation of these processes but are impractical at the field scale. Reduced complexity (e.g., field-scale or data-driven) models sacrifice some degree of fidelity in favor of computational efficiency. Uncertainty and assimilation of data into model predictions, pose a question of model selection: Given a significant difference in computational cost, can a lower-fidelity model be preferable to its higher-fidelity counterpart? Since a predictive model must be computable in reasonable time it has to operate at the field scale, with molecular- and pore-scale data (obtained either from simulations or measurements) determining both the model's structure and parameters. Availability of such multi-resolution data raises a question hitherto undressed in hydrogeology: Can a coarse model dynamically ``learn'' its own structure (e.g., adjust the reaction pathways in its transport module) as more fine-scale data become available during simulations? Our results in multifidelity simulations, data assimilation, and machine learning led us to hypothesize that the answer to these questions is ``yes''. The overarching goal of this project was to confirm this hypothesis by developing scalable, computationally efficient tools for assimilation of data into multiscale models, in which fine-scale data and simulations dynamically inform and autonomously modify coarse-scale models. Our research led to nine manuscripts, three of which have been published and the other six are currently under review.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

DOE OSTI2020

Compare linear-system solver and preconditioner stacks with emphasis on GPU performance and propose phase-2 NGP solver development pathway

The goal of the ExaWind project is to enable predictive simulations of wind farms comprised of many megawatt-scale turbines situated in complex terrain. Predictive simulations will require computational fluid dynamics (CFD) simulations for which the mesh resolves the geometry of the turbines and captures the rotation and large deflections of blades. Whereas such simulations for a single turbine are arguably petascale class, multi-turbine wind farm simulations will require exascale-class resources. The primary physics codes in the ExaWind project are Nalu-Wind, which is an unstructured-grid solver for the acoustically incompressible Navier-Stokes equations, and OpenFAST, which is a whole-turbine simulation code. The Nalu-Wind model consists of the mass-continuity Poisson-type equation for pressure and a momentum equation for the velocity. For such modeling approaches, simulation times are dominated by linear-system setup and solution for the continuity and momentum systems. For the ExaWind challenge problem, the moving meshes greatly affect overall solver costs as reinitialization of matrices and recomputation of preconditioners is required at every time step. In this report we evaluated GPU-performance baselines for the linear solvers in the Trilinos and hypre solver stacks using two representative Nalu-Wind simulations: an atmospheric boundary layer precursor simulation on a structured mesh, and a fixed-wing simulation using unstructured overset meshes. Both strong-scaling and weak-scaling experiments were conducted on the OLCF supercomputer Summit and similar proxy clusters. We focused on the performance of multi-threaded Gauss-Seidel and two-stage Gauss-Seidel that are extensions of classical Gauss-Seidel; of one-reduce GMRES, a communication-reducing variant of the Krylov GMRES; and algebraic multigrid methods that incorporate the afore-mentioned methods. The team has established that AMG methods are capable of solving linear systems arising from the fixed-wing overset meshes on CPU, a critical intermediate result for ExaWind FY20 Q3 and Q4 milestones. For the fixed-wing strong-scaling study (model with 3M grid-points), the team identified that Nalu-Wind simulations with the new Trilinos and hypre solvers scale to modest GPU counts, maintaining above 70% efficiency up to 6 GPUs. However, there still remain significant bottlenecks to performance: matrix assembly (hypre), AMG setup (hypre and Trilinos) In the weak-scaling experiments (going from 0.4M to 211M gridpoints), it's shown that the solver apply phases are faster on GPUs, but that Nalu-Wind simulation times grow, primarily due to the multigrid-setup process. Finally, based on the report outcomes, we propose a linear solver path-forward for the remainder of the ExaWind project. Near term, the NREL team will continue their work on GPU-based linear-system assembly. They will also investigate how the use of alternatives to the NVIDIA UVM (unified virtual memory) paradigm affects performance. Longer term, the NREL team will evaluate algorithmic performance on other types of accelerators and merge their improvements back to the main hypre repository branch. Near term, the Trilinos team will address performance bottlenecks identified in this milestone, such as implementing a GPU-based segregated momentum solve and reusing matrix graphs across linear-system assembly phases. Longer term, the Trilinos team will do detailed analysis and optimization of multigrid setup.

17 WIND ENERGY↗

DOE OSTI2020

Demonstration and performance testing of extreme-resolution simulations with static meshes on Summit (CPU & GPU) for a parked-turbine configuration and an actuator-line (mid-fidelity model) wind farm configuration (ECP-Q4 FY2020 Milestone Report)

17 WIND ENERGY↗

DOE OSTI2021

Robust Distributed State Estimator for Interconnected Transmission and Distribution Networks (Final Report RPPR-1)

This project’s objective is to develop a combined transmission and distribution state estimator which accounts for very large system size and model complexity (by way of distributing the computations) and large number of solar PV units connected to the distribution system on multiple feeders. The project not only provides a robust formulation and solution to this problem but also tests the solution by implementing it on a well-established large utility system. It introduces several improvements with respect to the state of the art in existing state estimation software: (a) The developed state estimator (SE) allows robust and accurate monitoring of bidirectional flows in distribution systems which result due to the distributed energy sources which are not observable and thus not incorporated in generation dispatch; (b) Large utility systems with tens of thousands of transmission buses and hundreds of thousands of distribution nodes are difficult to model as a single integrated system. This shortcoming is addressed by developing a “scalable distributed computational framework” which allows splitting the ultra large system models into several small subsystems and coordinating their solution by a robust and practical state estimation formulation; (c) Measurement errors irrespective of their locations are detected and removed by the developed state estimator. Historically, transmission and distribution systems were analyzed and operated as two independent systems. Given the non-transposed short feeder sections, unevenly loaded phases, strictly radial configuration and unidirectional power flows in the absence of remote generation, distribution system analysis was customized to account for these characteristics. However, some of these assumptions are no longer valid (non-radial configuration, bidirectional power flows) and thus distribution system analysis should be revisited. Furthermore, in the past, the interaction between the transmission and distribution systems was quite passive, where distribution substations were modeled as lumped loads in the transmission system model. With substantial generation injected by renewable generation located in the distribution systems, such modeling will no longer be accurate. The developed state estimator facilitates proper monitoring of the interactions between the transmission and distribution systems and enables smart dispatch of these units which are made observable by the state estimator.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2021

Vibration-Based Sensor Design: A Grey-Box Approach

Knowledge of the internal structure of an object or device under investigation proceeds from the basic idea of constructing its dynamic behavioral relations governed by a set of differential/algebraic equations that characterize its response. These equations can be partial differential equations leading to finite element or finite difference relations requiring a complex numerical solution on a super computer or ordinary differential equations requiring sophisticated numerical integration techniques to obtain the desired solution. Discrete dynamic systems evolving from digitized data acquisition are typically captured by sampled-data (continuous-to-discrete) representations characterized by a set of difference equations specifying the underlying system dynamics. In any case, with a mathematical description in hand, Grey-Box modeling techniques have evolved, concerned with the estimation of model parameters embedded in a prescribed set of equations (the system) governing its behavior, while capturing the underlying physical phenomenology of the problem at hand.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Demonstrate moving-grid multi-turbine simulations primarily run on GPUs and propose improvements for successful KPP-2

17 WIND ENERGY↗

DOE OSTI2022

Demonstrate multi-turbine simulation with hybrid-structured / unstructured-moving-grid software stack running primarily on GPUs and propose improvements for successful KPP-2

17 WIND ENERGY↗

DOE OSTI2023

Solutions for Enhanced Legacy Probabilistic Risk Assessment Tools and Methodologies: Improving Efficiency of Model Development and Processing via Innovative Human Reliability Dependency Analysis

Probabilistic risk assessments (PRAs) are integral to nuclear power plant (NPP) operations, having tremendously benefitted the safety of the U.S. reactor fleet for decades. Insights obtained from the models have provided perspectives on a variety of applications, both at the plant and for the regulator. While these models are very useful, they are now being asked to represent and analyze aspects of the plant that were never envisioned by the initial PRA practitioners. Furthermore, heightened demands on the PRA models have led to increased computing power requirements. Additionally, as the complexity of the PRA models increased, the difficulty experienced by non-PRA experts in trying to understand these models, grasp the insights they provide, and effectively use that information has become problematic. The need for research to address key issues regarding PRA tools and methods has never been greater. Although the nuclear power industry has largely been well-served by these tools and methods, the underlying science is dated, remaining mostly unchanged for over two decades. Three areas were identified as most beneficial to address to maintain and improve the usefulness of the current practice legacy PRA tools: improved quantification speed, increased ability to efficiently model multi-hazard models, and improved modeling human action dependency in PRA. This report is focused on the third critical area, improvements in dependency analysis of human actions conducted as part of a typical human reliability assessment.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2021

Efficient Learning of Accurate Surrogates for Simulations of Complex Systems

Machine learning methods are increasingly deployed to construct surrogate models for complex physical systems at a reduced computational cost. However, the predictive capability of these surrogates degrades in the presence of noisy, sparse, or dynamic data. We introduce an online learning method empowered by optimizer-driven sampling that has two advantages over current approaches: it ensures that all local extrema (including endpoints) of the model response surface are included in the training data and it employs a continuous validation and update process in which surrogates undergo retraining when their performance falls below a validity threshold. We find, using benchmark functions, that optimizer-directed sampling generally outperforms traditional sampling methods in terms of accuracy around local extrema even when the scoring metric is biased towards assessing overall accuracy. Finally, the application to dense nuclear matter demonstrates that highly accurate surrogates for a nuclear equation of state model can be reliably auto-generated from expensive calculations using few model evaluations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

AWAKEN Wind Plant Simulation Comparison

A series of numerical simulations of wind farms, using different model fidelities and for different atmospheric stability conditions, were performed as a part of the American WAKE ExperimeNt. The simulations included using FLORIS wake models, a number of microscale AMR-Wind and Nalu-Wind runs, as well as idealized and complex terrain WRF runs. The largest computations used the AMR-Wind LES solver to simulate a 100 km x 100 km domain containing 541 turbines under unstable atmospheric conditions matching previous measurements, while other LES computations focused on sections of the King Plains wind farm. Results of this qualitative comparison illustrate the interactions with wind farms with large-scale ABL structures in the flow, as well as the extent of downstream wake penetration in the flow and blockage effects around wind farms.

17 WIND ENERGY↗

DOE OSTI2026

Cavitand-Mediated Photodimerization of Chalcones: The Effect of Supramolecular Influences and Temperature on Reaction Selectivity

The photocycloaddition (PCA) of chalcones represents an important reaction pathway for accessing substituted cyclobutanes, which is a molecular framework with utility in synthetic chemistry, materials science, and medicine. In the past, our group has demonstrated the utility of the large cavity of γ-CD as a container for encapsulating two photo reactants for directing the PCA of several classes of aryl alkenes with high stereo- and regioselectivity: the cavitand-mediated photodimerization (CMP) approach. The CMP of chalcones reported in this work further demonstrates the effectiveness of this approach as high yields of dimers were observed in the photoreactions, while they were non-reactive in the solid state and yielded only the isomerization product in homogeneous media. The γ-CD CMP of chalcones yielded predominantly dimerized products in very good to high yields (>70%), composed of a mixture of three dimers in different proportions with syn HH as the major product. Computational analysis of the ground state complex structures revealed a strong correlation between the stability of the complex and predominance of the stereoisomer in the mixture. Further insights were deduced from temperature-dependence studies, which showed a shift in dimer selectivity tending towards a single stereoisomer.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2023

Oscillator for adiabatic computational circuitry

An adiabatic resonator, an adiabatic oscillator, and an adiabatic oscillator system are disclosed. An adiabatic system is one that ideally transfers no heat outside of the system, thereby reducing the required operating power. The adiabatic resonator, which includes a plurality of tank circuits, acts as an energy reservoir, the missing aspect of previously attempted adiabatic computational systems. By using the adiabatic resonator as a feedback element with an amplifier, an adiabatic oscillator is formed. An adiabatic oscillator system is formed with a primary adiabatic oscillator feeding a plurality of secondary adiabatic oscillators. In this manner, the adiabatic oscillator system may be used to generate the multiple clock signals required of adiabatic computational logic elements, such as Split-level Charge Recovery Logic and 2-Level Adiabatic Logic. The adiabatic oscillator system stores enough energy to drive many individual adiabatic computational logic elements, permitting implementation of complex logic circuits.

Brocato, Robert W.↗

DOE OSTI2024

Gas-Phase Coordination of Phosphine-Chalcogenides to the Uranyl Cation

It is generally believed that actinides exhibit an increased covalency in bonding to soft-donor atoms such as nitrogen and sulfur compared to the lanthanides. The explanation is that the greater spatial extent of the 5f orbitals in the actinides compared to the 4f orbitals in the lanthanides accounts for this behavior. However, recent computational studies on lanthanide and actinide complexes with sulfur-donor atom ligands suggest that while calculated metrics of bond covalency could increase for actinide complexes with sulfur ligands, the overall stability of the complexes decreased. While many studies of actinide covalency were conducted in the condensed phases, this work investigates the intrinsic interaction between actinides and soft-donor ligands compared with oxygen in the gas phase.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗