Machine Learning for Turbulence
New Deep Learning architectures that embed physics as hard constraints with numerical methods, increasing accuracy and reducing computational cost.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
SEARCH · Engineering Papers
Results for “computational costs”
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.
Kernel fusion in atomistic spin dynamics simulations on Nvidia GPUs using tensor core
In atomistic spin dynamics simulations, the time cost of constructing the space- and time-displaced pair correlation function in real space increases quadratically as the number of spins N, leading to significant computational effort. The GEMM subroutine can be adopted to accelerate the calculation of the dynamical spin-spin correlation function, but the computational cost of simulating large spin systems (>40000 spins) on CPUs remains expensive. In this work, we perform the simulation on the graphics processing unit (GPU), a hardware solution widely used as an accelerator for scientific computing and deep learning. Here we show that GPUs can accelerate the simulation up to 25-fold compared to multi-core CPUs when using the GEMM subroutine on both. To hide memory latency, we fuse the element-wise operation into the GEMM kernel using CUTLASS that can improve the performance by 26% ~ 33% compared to implementation based on cuBLAS. Furthermore, we perform the on-the-fly calculation in the epilogue of the GEMM subroutine to avoid saving intermediate results on global memory, which makes the large-scale atomistic spin dynamics simulation feasible and affordable.
97 MATHEMATICS AND COMPUTING↗
Dominant balance-based adaptive mesh refinement for incompressible fluid flows
This work introduces a novel adaptive mesh refinement (AMR) method that utilizes dominant balance analysis (DBA) for efficient and accurate grid adaptation in computational fluid dynamics (CFD) simulations. The proposed method leverages a Gaussian mixture model (GMM) to classify grid cells into active and passive regions based on the dominant physical interactions within the equation space. By modeling truncation error probabilistically from discretized terms, the method identifies regions of high interaction where numerical accuracy is most sensitive to resolution. Unlike traditional AMR strategies, this approach does not rely on heuristic-based sensors or user-defined thresholds, providing a fully automated and problem-independent framework for AMR. Applied to the incompressible Navier-Stokes equations for steady and unsteady flow past a cylinder, the DBA-based AMR method achieves comparable accuracy to high-resolution grids while reducing computational costs by up to 70 %. The validation highlights the method’s effectiveness in capturing complex flow features while minimizing grid cells, directing computational resources toward regions with the most critical dynamics. This modular and scalable strategy is adaptable to a wide range of applications, presenting a promising tool for efficient high-fidelity simulations in CFD and other multiphysics domains.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
Roadmap for unconventional computing with nanotechnology
Abstract In the ‘Beyond Moore’s Law’ era, with increasing edge intelligence, domain-specific computing embracing unconventional approaches will become increasingly prevalent. At the same time, adopting a variety of nanotechnologies will offer benefits in energy cost, computational speed, reduced footprint, cyber resilience, and processing power. The time is ripe for a roadmap for unconventional computing with nanotechnologies to guide future research, and this collection aims to fill that need. The authors provide a comprehensive roadmap for neuromorphic computing using electron spins, memristive devices, two-dimensional nanomaterials, nanomagnets, and various dynamical systems. They also address other paradigms such as Ising machines, Bayesian inference engines, probabilistic computing with p-bits, processing in memory, quantum memories and algorithms, computing with skyrmions and spin waves, and brain-inspired computing for incremental learning and problem-solving in severely resource-constrained environments. These approaches have advantages over traditional Boolean computing based on von Neumann architecture. As the computational requirements for artificial intelligence grow 50 times faster than Moore’s Law for electronics, more unconventional approaches to computing and signal processing will appear on the horizon, and this roadmap will help identify future needs and challenges. In a very fertile field, experts in the field aim to present some of the dominant and most promising technologies for unconventional computing that will be around for some time to come. Within a holistic approach, the goal is to provide pathways for solidifying the field and guiding future impactful discoveries.
Finocchio, Giovanni (ORCID:0000000210433876)↗
PREEMPT: Scalable Epidemic Interventions Using Submodular Optimization on Multi-GPU Systems
Preventing and slowing the spread of epidemics is achieved through techniques such as vaccination and social distancing. Given practical limitations on the number of vaccines and cost of administration, optimization becomes a necessity. Previous approaches using mathematical programming methods have shown to be effective but are limited by computational costs. In this work, we make several contributions: First, we present a new approach for intervention via maximizing the influence of vaccinated nodes on the network. We call this method \preempt. Next, we prove submodular properties associated with the objective function of our method so that it aids in construction of an efficient greedy approximation strategy. Consequently, we present a new parallel algorithm based on greedy hill climbing for \preempt, and present an efficient parallel implementation for distributed CPU-GPU heterogeneous platforms. Our results demonstrate that \preempt{} is able to achieve a significant reduction (up to 6.75$\times$) in the percentage of people infected on a city-scale network. We also show strong scaling results of \preempt{} on 128 nodes of the Summit supercomputer. Our parallel implementation is able to significantly reduce time to solution, from hours to minutes on large networks. This work represents a first-of-its-kind effort in parallelizing greedy hill climbing and applying it toward devising effective interventions for epidemics.
Minutoli, Marco↗
On differentiable local bounds preserving stabilization for Euler equations
This work presents the design of nonlinear stabilization techniques for the finite element discretization of Euler equations in both steady and transient form. Implicit time integration is used in the case of the transient form. A differentiable local bounds preserving method has been developed, which combines a Rusanov artificial diffusion operator and a differentiable shock detector. Nonlinear stabilization schemes are usually stiff and highly nonlinear. This issue is mitigated by the differentiability properties of the proposed method. Moreover, in order to further improve the nonlinear convergence, we also propose a continuation method for a subset of the stabilization parameters. The resulting method has been successfully applied to steady and transient problems with complex shock patterns. Numerical experiments show that it is able to provide sharp and well resolved shocks. Furthermore, the importance of the differentiability is assessed by comparing the new scheme with its non-differentiable counterpart. Numerical experiments suggest that, for up to moderate nonlinear tolerances, the method exhibits improved robustness and nonlinear convergence behavior for steady problems. Additionally, in the case of transient problem, we also observe a reduction in the computational cost.
42 ENGINEERING↗
An Infinite Domain 3D Poisson Solver Based on the Barnes-Hut Algorithm
We present a domain decomposition method for the solution of the 3D Poisson equation with infinite domain boundary conditions. The method is based on an application of the Barnes-Hut tree particle scheme adapted to gridded data. Long range interactions are computed using the first two terms in the Cartesian multipole expansion of Green’s function convoluted with the charge while short range computations are performed using Hockney’s domain doubling algorithm. A standard domain decomposition strategy requires O(N 2 ) applications of Hockney’s algorithm, where N is the number of subdomains that intersect that charge support, while in the present approach only O(Nlog 2 N) such computations suffice. The discretization scheme employed is a sixth order Mehrstellen approximation of the 3D Laplace opera tor. The method exhibits satisfactory accuracy at a substantially reduced computational cost compared to the full domain decomposition Hockney’s algorithm.
97 MATHEMATICS AND COMPUTING↗
aphBO-2GP-3B: a budgeted asynchronous parallel multi-acquisition functions for constrained Bayesian optimization on high-performing computing architecture
High-fidelity complex engineering simulations are often predictive, but also computationally expensive and often require substantial computational efforts. The mitigation of computational burden is usually enabled through parallelism in high-performance cluster (HPC) architecture. Optimization problems associated with these applications is a challenging problem due to the high computational cost of the high-fidelity simulations. In this paper, an asynchronous parallel constrained Bayesian optimization method is proposed to efficiently solve the computationally expensive simulation-based optimization problems on the HPC platform, with a budgeted computational resource, where the maximum number of simulations is a constant. The advantage of this method are three-fold. Firstly, the efficiency of the Bayesian optimization is improved, where multiple input locations are evaluated parallel in an asynchronous manner to accelerate the optimization convergence with respect to physical runtime. This efficiency feature is further improved so that when each of the inputs is finished, another input is queried without waiting for the whole batch to complete. Second, the proposed method can handle both known and unknown constraints. Third, the proposed method samples several acquisition functions based on their rewards using a modified GP-Hedge scheme. The proposed framework is termed aphBO-2GP-3B, which means asynchronous parallel hedge Bayesian optimization with two Gaussian processes and three batches. The numerical performance of the proposed framework aphBO-2GP-3B is comprehensively benchmarked using 16 numerical examples, compared against other 6 parallel Bayesian optimization variants and 1 parallel Monte Carlo as a baseline, and demonstrated using two real-world high-fidelity expensive industrial applications. The first engineering application is based on finite element analysis (FEA) and the second one is based on computational fluid dynamics (CFD) simulations.
97 MATHEMATICS AND COMPUTING↗
Multi-site reaction dynamics through multi-fragment density matrix embedding
The practical description of disordered chemical reactions, where the reactions involve multiple species at multiple sites, is presently a challenge using correlated electronic structure methods due to their high computational cost and steep scaling. Here, we describe the gradient theory of multi-fragment density matrix embedding theory, which potentially provides a minimal computational framework to model such processes at the correlated electron level. We present the derivation and implementation of the gradient theory, its validation on model systems and chemical reactions using density matrix embedding, and its application to a molecular dynamics simulation of proton transport in a small water cluster, a simple example of multi-site reaction dynamics.
Chemistry↗
A kinetic line-driven radiation operator and its application to Gyrokinetics
A velocity dependent, kinetic model for line radiation is developed for continuum kinetic codes. It has been implemented in the full-f gyrokinetic code Gkeyll. The total radiation for a charge state is modeled as an advection in velocity space with a form of $\nabla_v \cdot(v\nu(v)f(v))$, guaranteeing particle conservation. The velocity dependence (in the form of an effective frequency $\nu(v)$) is found through fitting the energy loss of the operator, i.e. the second velocity moment, to the radiation data in the OpenADAS database. Therefore, each individual transition does not need to be evaluated every time step, significantly reducing the computational cost of including line radiation in a kinetic model. The dependence on velocity instead of the usual, temperature, allows the radiation to be computed from non-Maxwellian electron distribution functions: We benchmark the model against a collisional radiative model using isotropic non-Maxwellian distribution functions. A velocity dependent model of radiation can more accurately describe the radiation in the more kinetic regimes expected in reactor-scale devices. The velocity dependence qualitatively captures the quantum mechanical need for a minimum velocity before any radiation occurs.
kinetic↗
Data and scripts associated with the manuscript "Using Mutual Information for Global Sensitivity Analysis on Watershed Modeling"
This package contains the data and scripts used in "Using Mutual Information for Global Sensitivity Analysis on Watershed Modeling" (Jiang et al., 2022). The ARW_SWAT.zip file contains the SWAT simulation and the sensitivity analysis result in the American River Watershed. The Portage_SWAT.zip file contains the SWAT simulation and the sensitivity analysis result in the Portage River Watershed. The src folder contains the source code of sensitivity analysis and plotting functions. The notebooks folder contains the Jupyter notebooks for performing sensitivity analysis and other post analyses.Global Sensitivity Analysis (GSA) often is applied to assess the sensitivity of model outputs to their inputs using ensemble simulations. However, increasing model complexity and the associated computational cost have limited the use of most GSA approaches for process-based watershed models. We propose to use Mutual Information (MI) as a computationally efficient GSA method for watershed modeling. Such MI computed from several hundred realizations usually can capture nonlinear relationships between inputs and outputs of interest. We perform MI-based watershed sensitivity analyses in studies of the Portage River Watershed in Ohio and the American River Watershed in Washington. In these studies, MI is used to evaluate the sensitivity of river discharges simulated by the Soil and Water Assessment Tool to no less than 20 parameters for each watershed. Our MI-based sensitivity analyses achieved convergence with about 300~500 realizations, a small fraction of the ensemble size (i.e., several thousand) required by the Sobol method. Nevertheless, MI yields similar sensitivity ranking compared to the Sobol method, especially for sensitive parameters. Our study thus sheds new light on the use of MI as an affordable GSA method for computationally intensive models such as the hyper-resolution, watershed hydrobiogeochemical models.
54 ENVIRONMENTAL SCIENCES↗
The curled wake model: a three-dimensional and extremely fast steady-state wake solver for wind plant flows
Abstract. Wind turbine wake models typically require approximations, such as wake superposition and deflection models, to accurately describe wake physics. However, capturing the phenomena of interest, such as the curled wake and interaction of multiple wakes, in wind power plant flows comes with an increased computational cost. To address this, we propose a new hybrid method that uses analytical solutions with an approximate form of the Reynolds-averaged Navier–Stokes equations to solve the time-averaged flow over a wind plant. We compare results from the solver to supervisory control and data acquisition data from the Lillgrund wind plant obtaining wake model predictions which are generally within 1 standard deviation of the mean power data. We perform simulations of flow over the Columbia River Gorge to demonstrate the capabilities of the model in complex terrain. We also apply the solver to a case with wake steering, which agreed well with large-eddy simulations. This new solver reduces the time – and therefore the related cost – it takes to simulate a steady-state wind plant flow (on the order of seconds using one core). Because the model is computationally efficient, it can also be used for different applications including wake steering for wind power plants and layout optimization.
17 WIND ENERGY↗
The Adjoint Petrov–Galerkin method for non-linear model reduction
Here, we formulate a new projection-based reduced-order modeling technique for non-linear dynamical systems. The proposed technique, which we refer to as the Adjoint Petrov–Galerkin (APG) method, is derived by decomposing the generalized coordinates of a dynamical system into a resolved coarse-scale set and an unresolved fine-scale set. A Markovian finite memory assumption within the Mori–Zwanzig formalism is then used to develop a reduced-order representation of the coarse scales. This procedure leads to a closed reduced-order model that displays commonalities with the adjoint stabilization method used in finite elements. The formulation is shown to be equivalent to a Petrov–Galerkin method with a non-linear, time-varying test basis, thus sharing some similarities with the Least-Squares Petrov–Galerkin method. Theoretical analysis examining a priori error bounds and computational cost is presented. Numerical experiments on the compressible Navier–Stokes equations demonstrate that the proposed method can lead to improvements in numerical accuracy, robustness, and computational efficiency over the Galerkin method on problems of practical interest. Improvements in numerical accuracy and computational efficiency over the Least-Squares Petrov–Galerkin method are observed in most cases.
42 ENGINEERING↗
Dual-cone variational calculation of the two-electron reduced density matrix
The computation of strongly correlated quantum systems is challenging because of its potentially exponential scaling in the number of electron configurations. Variational calculation of the two-electron reduced density matrix (2-RDM) without the many-electron wave function exploits the pairwise nature of the electronic Coulomb interaction to compute a lower bound on the ground-state energy with polynomial computational scaling. Recently, a dual-cone formulation of the variational 2-RDM calculation was shown to generate the ground-state energy, albeit not the 2-RDM, at a substantially reduced computational cost, especially for higher $\textit{N}$-representability conditions such as the T2 constraint. Here we generalize the dual-cone variational 2-RDM method to compute not only the ground-state energy but also the 2-RDM. The central result is that we can compute the 2-RDM from a generalization of the Hellmann-Feynman theorem. Specifically, we prove that in the Lagrangian formulation of the dual-cone optimization the 2-RDM is the Lagrange multiplier. Further, we apply the method to computing the energies and properties of strongly correlated electrons—including atomic charges, electron densities, dipole moments, and orbital occupations—in an illustrative hydrogen chain and the nitrogen-fixation catalyst FeMoco. The dual variational computation of the 2-RDM with T2 or higher N -representability conditions provides a polynomially scaling approach to strongly correlated molecules and materials with significant applications in atomic and molecular and condensed-matter chemistry and physics.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
Improving Trustworthiness of Data-Driven Power Grid Contingency Analysis With Bayesian Residual Graph Neural Networks
The evolving energy landscape requires novel tools to efficiently perform contingency analysis and reliability assessment of power grids, potentially in real-time. The high computational cost of traditional power flow solvers limits their applicability in practice. Machine learning (ML) surrogates such as deep neural networks (NNs) accelerate power flow solvers computations, enabling high-order contingency analysis and real-time decision-making by learning highly nonlinear functions and integrating grid topology via graph architectures. However, (graph) NNs lack predictive power away from training data and do not provide predictive confidence estimates. Here, we present a Bayesian residual graph NN that integrates knowledge from low-fidelity data via residual training and embeds granular quantification of uncertainties, improving trustworthiness critical for high-consequence decision-making. Applying Bayesian concepts to NNs is challenging due to the high-dimensionality of both the parameter space, complicating derivation of a meaningful prior, and the output space in large grid systems, requiring enhanced techniques to assess the predicted high-dimensional uncertainties. Our contributions include: (1) Deriving a prior for fully connected and graph NNs that leverages low-fidelity data to guide mean predictions and appropriately control prior predictive uncertainty. (2) Integrating this prior within an ensembling with anchoring scheme for efficient approximate posterior inference. (3) Deriving enhanced metrics to assess accuracy of both the mean and uncertainty predictions in high dimensions, appropriately accounting for correlations propagated through graph layers. The resulting Bayesian residual graph NN is tested on a contingency analysis task for 14-bus and 118-bus grids.
24 - POWER TRANSMISSION AND DISTRIBUTION↗
A priori examination of reduced chemistry models derived from canonical stirred reactors using three-dimensional direct numerical simulation datasets
Data-driven approaches to construct reduced chemical kinetic models, that rely heavily on thermo-chemical datasets with full chemical kinetics, have been gaining popularity. Datasets from direct numerical simulations (DNS) under three-dimensional (3-D) realistic turbulent flow conditions are desirable but limited to carefully designed parametric conditions due to the computational cost. Constructing datasets from a large ensemble of zero-dimensional stirred reactors like perfectly stirred reactor (PSR) and partially stirred reactor (PaSR) is a computationally efficient solution to consider the turbulence-chemistry interactions and cover a broad range of parametric conditions. In this paper, we derive reduced chemistry models from solutions of a large number of PSR and PaSR reactors using autoencoder (AE) neural networks and principal component analysis (PCA), and conduct a priori examination of the reduced models in three temporally evolving 3-D DNS jet flames featuring local extinction and re-ignition. The results show that the reduced models derived from PaSR datasets, i.e., AE-PaSR and PCA-PaSR, generally show significant improvement over the ones derived from PSR datasets. Among all the reduced models, AE-PaSR shows the best agreement with DNS results on the reconstruction accuracy and the representation of temporally evolving local extinction and re-ignition events.
Zhang, Pei↗
A Systematic, Polynomial-Cost Approach to Exact Correlation Energies (Final Technical Report)
The full configuration interaction (FCI) wave function provides the exact solution to the Schrödinger equation in a given basis set. While FCI is intractable due to exponential computational costs, the many-body expansion (or the method of increments) can reduce scaling to a low-order polynomial with system size. This project entailed advances in the incremental FCI (iFCI) approach, designed to allow iFCI to reach larger system sizes and maintain its intrinsic high accuracy. This document describes advances in solvers for iFCI, strategies to treat multiple charge and spin states, and virtual state management methods to reduce memory requirements. Overall, this project allows iFCI to correlate (for the first time) 142 valence electrons in 444 orbitals in a realistic model of a transition metal complex.
74 ATOMIC AND MOLECULAR PHYSICS↗
Tula: Optimizing Time, Cost, and Generalization in Distributed Large-Batch Training
Distributed training increases the number of batches processed per iteration either by scaling-out (adding more nodes) or scaling-up (increasing the batch-size). However, the largest configuration does not necessarily yield the best performance. Horizontal scaling introduces additional communication overhead, while vertical scaling is constrained by computation cost and device memory limits. Thus, simply increasing the batch-size leads to diminishing returns: training time and cost decrease initially but eventually plateaus, creating a knee-point in the time/cost vs. batch-size pareto curve. The optimal batch-size therefore depends on the underlying model, data and available compute resources. Large batches also suffer from worse model quality due to the well-known “generalization gap”. In this paper, we present Tula, an online service that automatically optimizes time, cost, and convergence quality for large-batch training of convolutional models. It combines parallel-systems modeling with statistical performance prediction to identify the optimal batchsize. Tula predicts training time and cost within 7.5−14% error across multiple models, and achieves up to 20× overall speedup and improves test accuracy by ≈9% on average over standard large-batch training on various vision tasks, thus successfully mitigating the generalization gap and accelerating training at the same time.
Tyagi, Sahil [ORNL] (ORCID:0009000783144745)↗