Sondak, David; Smith, Thomas M.; Pawlowski, Roger P.; Conde, Sidafa C.; Shadid, John N.
The variational multiscale (VMS) formulation is used to develop residual-based VMS large eddy simulation (LES) models for Rayleigh-Bénard convection. The resulting model is a mixed model that incorporates the VMS model and an eddy viscosity model. The Wall-Adapting Local Eddy-viscosity (WALE) model is used as the eddy viscosity model in this work. The new LES models were implemented in the finite element code Drekar. Simulations are performed using continuous, piecewise linear finite elements. The simulations ranged from Ra=106 to Ra=1014 and were conducted at Pr=1 and Pr=7. Two domains were considered: a two-dimensional domain of aspect ratio 2 with a fluid confined between two parallel plates and a three-dimensional cylinder of aspect ratio 1/4. The Nusselt number from the VMS results is compared against three dimensional direct numerical simulations and experiments. In all cases, the VMS results are in good agreement with existing literature.
The explosion of both sensors and GPS-enabled devices has resulted in position/time data being the next big frontier for data analytics. However, many of the problems associated with large numbers of trajectories do not necessarily have an analog with many of the historic big-data applications such as text and image analysis. Modern trajectory analytics exploits much of the cutting-edge research in machine-learning, statistics, computational geometry and other disciplines. We will show that for doing trajectory analytics at scale, it is necessary to fundamentally change the way the information is represented through a feature-vector approach. We then demonstrate the ability to solve large trajectory analytics problems using this representation.
The modern scientific process often involves the development of a predictive computational model. To improve its accuracy, a computational model can be calibrated to a set of experimental data. A variety of validation metrics can be used to quantify this process. Some of these metrics have direct physical interpretations and a history of use, while others, especially those for probabilistic data, are more difficult to interpret. In this work, a variety of validation metrics are used to quantify the accuracy of different calibration methods. Frequentist and Bayesian perspectives are used with both fixed effects and mixed-effects statistical models. Through a quantitative comparison of the resulting distributions, the most accurate calibration method can be selected. Two examples are included which compare the results of various validation metrics for different calibration methods. It is quantitatively shown that, in the presence of significant laboratory biases, a fixed effects calibration is significantly less accurate than a mixed-effects calibration. This is because the mixed-effects statistical model better characterizes the underlying parameter distributions than the fixed effects model. The results suggest that validation metrics can be used to select the most accurate calibration model for a particular empirical model with corresponding experimental data.
A complete inelastic equation of state (IEOS) for solids is developed based on a superposition of thermodynamic energy potentials. The IEOS allows for a tensorial stress state by including an isochoric hyperelastic Helmholtz potential in addition to the zero-kelvin isotherm and lattice vibration energy contributions. Inelasticity is introduced through the nonlinear equations of finite strain plasticity which utilize the temperature dependent Johnson–Cook yield model. Material failure is incorporated into the model by a coupling of the damage history variable to the energy potentials. The numerical evaluation of the IEOS requires a nonlinear solution of stress, temperature and history variables associated with elastic trial states for stress and temperature. The model is implemented into the ALEGRA shock and multi-physics code and the applications presented include single element deformation paths, the Taylor anvil problem and an energetically driven thermo-mechanical problem.
Peng, Ivy P.; Voskuilen, Gwendolyn R.; Sarkar, Abhik S.; Boehme, David B.; Long, Rogelio L.; Moore, Shirley M.; Gokhale, Maya G.
This is the second in a sequence of three Hardware Evaluation milestones that provide insight into the following questions: What are the sources of excess data movement across all levels of the memory hierarchy, going out to the network fabric? What can be done at various levels of the hardware/software hierarchy to reduce excess data movement? How does reduced data movement track application performance? The results of this study can be used to suggest where the DOE supercomputing facilities, working with their hardware vendors, can optimize aspects of the system to reduce excess data movement. Quantitative analysis will also benefit systems software and applications to optimize caching and data layout strategies. Another potential avenue is to answer cost-benefit questions, such as those involving memory capacity versus latency and bandwidth. This milestone focuses on techniques to reduce data movement, quantitatively evaluates the efficacy of the techniques in accomplishing that goal, and measures how performance tracks data movement reduction. We study a small collection of benchmarks and proxy mini-apps that run on pre-exascale GPUs and on the Accelsim GPU simulator. Our approach has two thrusts: to measure advanced data movement reduction directives and techniques on the newest available GPUs, and to evaluate our benchmark set on simulated GPUs configured with architectural refinements to reduce data movement.
This report focuses on the two primary goals set forth in Sandia’s TAFI effort, referred to here under the name Kebab. The first goal is to overlay a trajectory onto a large database of historical trajectories, all with very different sampling rates than the original track. We demonstrate a fast method to accomplish this, even for databases that hold over a million tracks. The second goal is to then demonstrate that these matched historical trajectories can be used to make predictions about unknown qualities associated with the original trajectory. As part of this work, we also examine the problem of defining the qualities of a trajectory in a reproducible way.
Garg, Raveesh G.; Qin, Eric Q.; Martinez, Francisco M.; Guirado, Robert G.; Jain, Akshay J.; Abadal, Sergi A.; Abellan, Jose L.; Acacio, Manuel E.; Alarcon, Eduard A.; Rajamanickam, Sivasankaran R.; Krishna, Tushar K.
Recently, Graph Neural Networks (GNNs) have received a lot of interest because of their success in learning representations from graph structured data. However, GNNs exhibit different compute and memory characteristics compared to traditional Deep Neural Networks (DNNs). Graph convolutions require feature aggregations from neighboring nodes (known as the aggregation phase), which leads to highly irregular data accesses. GNNs also have a very regular compute phase that can be broken down to matrix multiplications (known as the combination phase). All recently proposed GNN accelerators utilize different dataflows and microarchitecture optimizations for these two phases. Different communication strategies between the two phases have been also used. However, as more custom GNN accelerators are proposed, the harder it is to qualitatively classify them and quantitatively contrast them. In this work, we present a taxonomy to describe several diverse dataflows for running GNN inference on accelerators. This provides a structured way to describe and compare the design-space of GNN accelerators.
Lysogorskiy, Yury L.; Rinaldi, Matteo R.; Menon, Sarath M.; van der Oord, Cas v.; Hammerschmidt, Thomas H.; Mrovec, Matous M.; Thompson, Aidan P.; Csanyi, Gabor C.; Ortner, Christoph O.; Drautz, Ralf D.
The atomic cluster expansion is a general polynomial expansion of the atomic energy in multi-atom basis functions. Here we implement the atomic cluster expansion in the performant C++ code PACE that is suitable for use in large scale atomistic simulations. We briefly review the atomic cluster expansion and give detailed expressions for energies and forces as well as efficient algorithms for their evaluation. We demonstrate that the atomic cluster expansion as implemented in PACE shifts a previously established Pareto front for machine learning interatomic potentials towards faster and more accurate calculations. Moreover, general purpose parameterizations are presented for copper and silicon and evaluated in detail. We show that the new Cu and Si potentials significantly improve on the best available potentials for highly accurate large-scale atomistic simulations.
We present a fully discrete approximation technique for the compressible Navier–Stokes equations that is second-order accurate in time and space, semi-implicit, and guaranteed to be invariant domain preserving. The restriction on the time step is the standard hyperbolic CFL condition, i.e. τ≲O(h)∕V where V is some reference velocity scale and h the typical meshsize.
Learning 3D representations that generalize well to arbitrarily oriented inputs is a challenge of practical importance in applications varying from computer vision to physics and chemistry. We propose a novel multi-resolution convolutional architecture for learning over concentric spherical feature maps, of which the single sphere representation is a special case. Our hierarchical architecture is based on alternatively learning to incorporate both intra-sphere and inter-sphere information. We show the applicability of our method for two different types of 3D inputs, mesh objects, which can be regularly sampled, and point clouds, which are irregularly distributed. We also propose an efficient mapping of point clouds to concentric spherical images, thereby bridging spherical convolutions on grids with general point clouds. We demonstrate the effectiveness of our approach in improving state-of-the-art performance on 3D classification tasks with rotated data.
We study both conforming and non-conforming versions of the practical DPG method for the convection-reaction problem. We determine that the most common approach for DPG stability analysis (construction of a local Fortin operator) is infeasible for the convection-reaction problem. We then develop a line of argument based on the direct construction of a global Fortin operator; we find that employing a polynomial enrichment for the test space does not suffice for this purpose, motivating the introduction of a (two-element) subgrid mesh. The argument combines mathematical analysis with numerical experiments
ROL-PEBBL is a C++, MPI-based parallel code for mixed-integer PDE-constrained optimization (MIPDECO). In these problems we wish to optimize (control, design, etc.) physical systems, which must obey the laws of physics, when some of the decision variables must take integer values. ROL-PEBBL combines a code to efficiently search over integer choices (PEBBL = Parallel Enumeration Branch-and-Bound Library) and a code for efficient nonlinear optimization, including PDE-constrained optimization (ROL = Rapid Optimization Library). In this report, we summarize the design of ROL-PEBBL and initial applications/results. For an artificial source-inversion problem, finding sources of pollution on a grid from sparse samples, ROL-PEBBLs solution for the nest grid gave the best optimization guarantee for any general solver that gives both a solution and a quality guarantee.
In this position paper we will address challenges and opportunities relating to the design and codesign of application specific circuits. Given our background as computational scientists, our perspective is from the viewpoint of a highly motivated application developer as opposed to career computer architects
Hart, Joseph L.; Gilmore, Steven G.; Gremaud, Pierre G.; Olsen, Christian O.; Mehlsen, Jesper M.; Olufsen, Mette O.
Imbalance in the autonomic nervous system can lead to orthostatic intolerance manifested by dizziness, lightheadedness, and a sudden loss of consciousness (syncope); these are common conditions, but they are challenging to diagnose correctly. Uncertainties about the triggering mechanisms and the underlying pathophysiology have led to variations in their classification. This study uses machine learning to categorize patients with orthostatic intolerance. Here we use random forest classification trees to identify a small number of markers in blood pressure, and heart rate time-series data measured during head-up tilt to (a) distinguish patients with a single pathology and (b) examine data from patients with a mixed pathophysiology. Next, we use Kmeans to cluster the markers representing the time-series data. We apply the proposed method analyzing clinical data from 186 subjects identified as control or suffering from one of four conditions: postural orthostatic tachycardia (POTS), cardioinhibition, vasodepression, and mixed cardioinhibition and vasodepression. Classification results confirm the use of supervised machine learning. We were able to categorize more than 95% of patients with a single condition and were able to subgroup all patients with mixed cardioinhibitory and vasodepressor syncope. Clustering results confirm the disease groups and identify two distinct subgroups within the control and mixed groups. The proposed study demonstrates how to use machine learning to discover structure in blood pressure and heart rate time-series data. The methodology is used in classification of patients with orthostatic intolerance. Diagnosing orthostatic intolerance is challenging, and full characterization of the pathophysiological mechanisms remains a topic of ongoing research. This study provides a step toward leveraging machine learning to assist clinicians and researchers in addressing these challenges.
MLIR (Multi-Level Intermediate Representation), is an extensible compiler framework that supports high-level data structures and operation constructs. These higher-level code representations are particularly applicable to the artificial intelligence and machine learning (AI/ML) domain, allowing developers to more easily support upcoming heterogeneous AI/ML accelerators and develop flexible domain specific compilers/frameworks with higher-level intermediate representations (IRs) and advanced compiler optimizations. The result of using MLIR within the LLVM compiler framework is expected to yield significant improvement in the quality of generated machine code, which in turn will result in improved performance and hardware efficiency
Non-volatile memories (NVMs) have the characteristics of both traditional storage systems (persistent) and traditional memory systems (byte-Addressable). However, they suffer from high write latency and have a limited write endurance. Researchers have proposed hybrid memory systems that combine DRAM and NVM, utilizing the lower latency of the DRAM to hide some of the shortcomings of the NVM-improving system's performance by caching resident NVM data in the DRAM. However, this can nullify the persistency of the cached pages, leading to a question of trade-offs in terms of performance and reliability. In this paper, we propose Stealth-Persist, a novel architecture support feature that allows applications that need persistence to run in the DRAM while maintaining the persistency features provided by the NVM. Stealth-Persist creates the illusion of a persistent memory for the application to use, while utilizing the DRAM for performance optimizations. Our experimental results show that Stealth-Persist improves the performance by 42.02% for persistent applications.
We present an optimization-based coupling method for local and nonlocal continuum models. Our approach couches the coupling of the models into a control problem where the states are the solutions of the nonlocal and local equations, the objective is to minimize their mismatch on the overlap of the local and nonlocal problem domains, and the virtual controls are the nonlocal volume constraint and the local boundary condition. We present the method in the context of Local-to-Nonlocal diffusion coupling. Numerical examples illustrate the theoretical properties of the approach.
1 The exponential growth of data has driven technology providers to develop new protocols, such as cache coherent interconnects and memory semantic fabrics, to help users and facilities leverage advances in memory technologies to satisfy these growing memory and storage demands. Using these new protocols, fabric-Attached memories (FAM) can be directly attached to a system interconnect and be easily integrated with a variety of processing elements (PEs). Moreover, systems that support FAM can be smoothly upgraded and allow multiple PEs to share the FAM memory pools using well-defined protocols. The sharing of FAM between PEs allows efficient data sharing, improves memory utilization, reduces cost by allowing flexible integration of different PEs and memory modules from several vendors, and makes it easier to upgrade the system. One promising use-case for FAMs is in High-Performance Compute (HPC) systems, where the underutilization of memory is a major challenge. However, adopting FAMs in HPC systems brings new challenges. In addition to cost, flexibility, and efficiency, one particular problem that requires rethinking is virtual memory support for security and performance. To address these challenges, this paper presents decoupled access control and address translation (DeACT), a novel virtual memory implementation that supports HPC systems equipped with FAM. Compared to the state-of-The-Art two-level translation approach, DeACT achieves speedup of up to 4.59x (1.8x on average) without compromising security.1Part of this work was done when Vamsee was working under the supervision of Amro Awad at UCF. Amro Awad is now with the ECE Department at NC State.
A key challenge to nonlocal models is the analytical complexity of deriving them from first principles, and frequently their use is justified a posteriori. In this work we extract nonlocal models from data, circumventing these challenges and providing data-driven justification for the resulting model form. Extracting data-driven surrogates is a major challenge for machine learning (ML) approaches, due to nonlinearities and lack of convexity — it is particularly challenging to extract surrogates which are provably well-posed and numerically stable. Our scheme not only yields a convex optimization problem, but also allows extraction of nonlocal models whose kernels may be partially negative while maintaining well-posedness even in small-data regimes. To achieve this, based on established nonlocal theory, we embed in our algorithm sufficient conditions on the non-positive part of the kernel that guarantee well-posedness of the learnt operator. These conditions are imposed as inequality constraints to meet the requisite conditions of the nonlocal theory. We demonstrate this workflow for a range of applications, including reproduction of manufactured nonlocal kernels; numerical homogenization of Darcy flow associated with a heterogeneous periodic microstructure; nonlocal approximation to high-order local transport phenomena; and approximation of globally supported fractional diffusion operators by truncated kernels.
For digital twins (DTs) to become a central fixture in mission critical systems, a better understanding is required of potential modes of failure, quantification of uncertainty, and the ability to explain a model’s behavior. These aspects are particularly important as the performance of a digital twin will evolve during model development and deployment for real-world operations.
Signal arrival-time estimation plays a critical role in a variety of downstream seismic analy-ses, including location estimation and source characterization. Any arrival-time errors propagate through subsequent data-processing results. In this article, we detail a general framework for refining estimated seismic signal arrival times along with full estimation of their associated uncertainty. Using the standard short-term average/long-term average threshold algorithm to identify a search window, we demonstrate how to refine the pick estimate through two different approaches. In both cases, new waveform realizations are generated through bootstrap algorithms to produce full a posteriori estimates of uncertainty of onset arrival time of the seismic signal. The onset arrival uncertainty estimates provide additional data-derived information from the signal and have the potential to influence seismic analysis along several fronts.
Over the past decade as Moore's Law has slowed, the need for new forms of computation that can provide sustainable performance improvements has risen. A new method, called in situ computing, has shown great potential to accelerate matrix vector multiplication (MVM), an important kernel for a diverse range of applications from neural networks to scientific computing. Existing in situ accelerators for scientific computing, however, have a significant limitation: These accelerators provide no acceleration for preconditioning-A key bottleneck in linear solvers and in scientific computing workflows. This paper enables in situ acceleration for state-of-The-Art linear solvers by demonstrating how to use a new in situ matrix inversion accelerator for analog preconditioning. As existing techniques that enable high precision and scalability for in situ MVM are inapplicable to in situ matrix inversion, new techniques to compensate for circuit non-idealities are proposed. Additionally, a new approach to bit slicing that enables splitting operands across multiple devices without external digital logic is proposed. For scalability, this paper demonstrates how in situ matrix inversion kernels can work in tandem with existing domain decomposition techniques to accelerate the solutions of arbitrarily large linear systems. The analog kernel can be directly integrated into existing preconditioning workflows, leveraging several well-optimized numerical linear algebra tools to improve the behavior of the circuit. The result is an analog preconditioner that is more effective (up to 50% fewer iterations) than the widely used incomplete LU factorization preconditioner, ILU(0), while also reducing the energy and execution time of each approximate solve operation by 1025x and 105x respectively.
n this presentation we will discuss recent results on using the SpiNNaker neuromorphic platform (48-chip model) for deep learning neural network inference. We use the Sandia Labs developed Whet stone spiking deep learning library to train deep multi-layer perceptrons and convolutional neural networks suitable for the spiking substrate on the neural hardware architecture. By using the massively parallel nature of SpiNNaker, we are able to achieve, under certain network topologies, substantial network tiling and consequentially impressive inference throughput. Such high-throughput systems may have eventual application in remote sensing applications where large images need to be chipped, scanned, and processed quickly. Additionally, we explore complex topologies that push the limits of the SpiNNaker routing hardware and investigate how that impacts mapping software-implemented networks to on-hardware instantiations.
The study of hypersonic flows and their underlying aerothermochemical reactions is particularly important in the design and analysis of vehicles exiting and reentering Earth's atmosphere. Computational physics codes can be employed to simulate these phenomena; however, verification of these codes is necessary to certify their credibility. To date, few approaches have been presented for verifying codes that simulate hypersonic flows, especially flows reacting in thermochemical nonequilibrium. In this paper, we present our code-verification techniques for verifying the spatial accuracy and thermochemical source term in hypersonic reacting flows in thermochemical nonequilibrium. We demonstrate the effectiveness of these techniques on the Sandia Parallel Aerodynamics and Reentry Code (SPARC).
Diborane (B2H6) is a promising molecular precursor for atomic precision p-type doping of silicon that has recently been experimentally demonstrated [ Škereň et al. Nat. Electron. 2020 ]. We use density functional theory (DFT) calculations to determine the reaction pathway for diborane dissociating into a species that will incorporate as electrically active substitutional boron after adsorbing onto the Si(100)-2×1 surface. Our calculations indicate that diborane must overcome an energy barrier to adsorb, explaining the experimentally observed low sticking coefficient (<1 × 10-4 at room temperature) and suggesting that heating can be used to increase the adsorption rate. Upon sticking, diborane has an ≈50% chance of splitting into two BH3 fragments versus merely losing hydrogen to form a dimer such as B2H4. As boron dimers are likely electrically inactive, whether this latter reaction occurs is shown to be predictive of the incorporation rate. The dissociation process proceeds with significant energy barriers, necessitating the use of high temperatures for incorporation. Using the barriers calculated from DFT, we parameterize a Kinetic Monte Carlo model that predicts the incorporation statistics of boron as a function of the initial depassivation geometry, dose, and anneal temperature. Our results suggest that the dimer nature of diborane inherently limits its doping density as an acceptor precursor and furthermore that heating the boron dimers to split before exposure to silicon can lead to poor selectivity on hydrogen and halogen resists. This suggests that, while diborane works as an atomic precision acceptor precursor, other non-dimerized acceptor precursors may lead to higher incorporation rates at lower temperatures.
Programmable accelerators have become commonplace in modern computing systems. Advances in programming models and the availability of unprecedented amounts of data have created a space for massively parallel accelerators capable of maintaining context for thousands of concurrent threads resident on-chip. These threads are grouped and interleaved on a cycle-by-cycle basis among several massively parallel computing cores. One path for the design of future supercomputers relies on an ability to model the performance of these massively parallel cores at scale. The SST framework has been proven to scale up to run simulations containing tens of thousands of nodes. A previous report described the initial integration of the open-source, execution-driven GPU simulator, GPGPU-Sim, into the SST framework. This report discusses the results of the integration and how to use the new GPU component in SST. It also provides examples of what it can be used to analyze and a correlation study showing how closely the execution matches that of a Nvidia V100 GPU when running kernels and mini-apps.
Persistent memory (PMEM) devices can achieve comparable performance to DRAM while providing significantly more capacity. This has made the technology compelling as an expansion to main memory. Rethinking PMEM as storage devices can offer a high performance buffering layer for HPC applications to temporarily, but safely store data. However, modern parallel I/O libraries, such as HDF5 and pNetCDF, are complicated and introduce significant software and metadata overheads when persisting data to these storage devices, wasting much of their potential. In this work, we explore the potential of PMEM as storage through pMEMCPY: a simple, lightweight, and portable I/O library for storing data in persistent memory. We demonstrate that our approach is up to 2x faster than other popular parallel I/O libraries under real workloads.
Rendezvous algorithms encode a communication pattern that is useful when processors sending data do not know who the receiving processors should be, or vice versa. The idea is to define an intermediate decomposition where datums from different sending processors can ”rendezvous” to perform a computation, in a manner that both the senders and eventual receivers of the results can identify the appropriate rendezvous processor. Originally designed for interpolating between overlaid grids with independent parallel decompositions (Plimpton et al., 2004), we have recently found rendezvous algorithms useful for a variety of operations in particle- or grid-based simulation codes when running large problems on large numbers of processors. In particular, we show they can perform well when a load-balanced intermediate decomposition is randomized and not spatial, requiring all-to-all communication to move data between processors. In this case rendezvous algorithms leverage the large bisection communication bandwidths which parallel machines provide. We describe how rendezvous algorithms work in a scientific computing context and give specific examples for molecular dynamics and Direct Simulation Monte Carlo codes which result in dramatic performance improvements versus simpler algorithms which do not scale as well. We explain how a generic rendezvous algorithm can be implemented, and also point out similarities with the MapReduce paradigm popularized by Google and Hadoop.
A class of sequential multiscale models investigated in this study consists of discrete dislocation dynamics (DDD) simulations and continuum strain gradient plasticity (SGP) models to simulate the size effect in plastic deformation of metallic micropillars. The high-fidelity DDD explicitly simulates the microstructural (dislocation) interactions. These simulations account for the effect of dislocation densities and their spatial distributions on plastic deformation. The continuum SGP captures the size-dependent plasticity in micropillars using two length parameters. The main challenge in predictive DDD-SGP multiscale modeling is selecting the proper constitutive relations for the SGP model, which is necessitated by the uncertainty in computational prediction due to DDD's microstructural randomness. This contribution addresses these challenges using a Bayesian learning and model selection framework. A family of SGP models with different fidelities and complexities is constructed using various constitutive relation assumptions. The parameters of the SGP models are then learned from a set of training data furnished by the DDD simulations of micropillars. Bayesian learning allows the assessment of the credibility of plastic deformation prediction by characterizing the microstructural variability and the uncertainty in training data. Additionally, the family of the possible SGP models is subjected to a Bayesian model selection to pick the model that adequately explains the DDD training data. The framework proposed in this study enables learning the physics-based multiscale model from uncertain observational data and determining the optimal computational model for predicting complex physical phenomena, i.e., size effect in plastic deformation of micropillars.
This report describes the high-level accomplishments from the Plasma Science and Engineering Grand Challenge LDRD at Sandia National Laboratories. The Laboratory has a need to demonstrate predictive capabilities to model plasma phenomena in order to rapidly accelerate engineering development in several mission areas. The purpose of this Grand Challenge LDRD was to advance the fundamental models, methods, and algorithms along with supporting electrode science foundation to enable a revolutionary shift towards predictive plasma engineering design principles. This project integrated the SNL knowledge base in computer science, plasma physics, materials science, applied mathematics, and relevant application engineering to establish new cross-laboratory collaborations on these topics. As an initial exemplar, this project focused efforts on improving multi-scale modeling capabilities that are utilized to predict the electrical power delivery on large-scale pulsed power accelerators. Specifically, this LDRD was structured into three primary research thrusts that, when integrated, enable complex simulations of these devices: (1) the exploration of multi-scale models describing the desorption of contaminants from pulsed power electrodes, (2) the development of improved algorithms and code technologies to treat the multi-physics phenomena required to predict device performance, and (3) the creation of a rigorous verification and validation infrastructure to evaluate the codes and models across a range of challenge problems. These components were integrated into initial demonstrations of the largest simulations of multi-level vacuum power flow completed to-date, executed on the leading HPC computing machines available in the NNSA complex today. These preliminary studies indicate relevant pulsed power engineering design simulations can now be completed in (of order) several days, a significant improvement over pre-LDRD levels of performance.
Second-order optimizers hold intriguing potential for deep learning, but suffer from increased cost and sensitivity to the non-convexity of the loss surface as compared to gradient-based approaches. We introduce a coordinate descent method to train deep neural networks for classification tasks that exploits global convexity of the cross-entropy loss in the weights of the linear layer. Our hybrid Newton/Gradient Descent (NGD) method is consistent with the interpretation of hidden layers as providing an adaptive basis and the linear layer as providing an optimal fit of the basis to data. By alternating between a second-order method to find globally optimal parameters for the linear layer and gradient descent to train the hidden layers, we ensure an optimal fit of the adaptive basis to data throughout training. The size of the Hessian in the second-order step scales only with the number weights in the linear layer and not the depth and width of the hidden layers; furthermore, the approach is applicable to arbitrary hidden layer architecture. Previous work applying this adaptive basis perspective to regression problems demonstrated significant improvements in accuracy at reduced training cost, and this work can be viewed as an extension of this approach to classification problems. We first prove that the resulting Hessian matrix is symmetric semi-definite, and that the Newton step realizes a global minimizer. By studying classification of manufactured two-dimensional point cloud data, we demonstrate both an improvement in validation error and a striking qualitative difference in the basis functions encoded in the hidden layer when trained using NGD. Application to image classification benchmarks for both dense and convolutional architectures reveals improved training accuracy, suggesting gains of second-order methods over gradient descent. A Tensorflow implementation of the algorithm is available at github.com/rgp62/.
A key strategy for protecting municipal water supplies is the use of sensors to detect the presence of contaminants in associated water distribution systems. Deploying a contamination warning system involves the placement of a limited number of sensors—placed in order to maximize the level of protection afforded. Researchers have proposed several models and algorithms for generating such placements, each optimizing with respect to a different design objective. The use of disparate design objectives raises several questions: (1) What is the relationship between optimal sensor placements for different design objectives? and (2) Is there any risk in focusing on specific design objectives? We model the sensor placement problem via a mixed-integer programming formulation of the well-known p-median problem from facility location theory to answer these questions. Our model can express a broad range of design objectives. Using three large test networks, we show that optimal solutions with respect to one design objective are often highly sub-optimal with respect to other design objectives. However, it is sometimes possible to construct solutions that are simultaneously near-optimal with respect to a range of design objectives. The design of contamination warning systems thus requires careful and simultaneous consideration of multiple, disparate design objectives.
35th AAAI Conference on Artificial Intelligence, AAAI 2021
Kim, Jungeun; Lee, Kookjin L.; Lee, Dongeun; Jhin, Sheo Y.; Park, Noseong
We present a method for learning dynamics of complex physical processes described by time-dependent nonlinear partial differential equations (PDEs). Our particular interest lies in extrapolating solutions in time beyond the range of temporal domain used in training. Our choice for a baseline method is physics-informed neural network (PINN) because the method parameterizes not only the solutions, but also the equations that describe the dynamics of physical processes. We demonstrate that PINN performs poorly on extrapolation tasks in many benchmark problems. To address this, we propose a novel method for better training PINN and demonstrate that our newly enhanced PINNs can accurately extrapolate solutions in time. Our method shows up to 72% smaller errors than existing methods in terms of the standard L2-norm metric.
Recent advances in neuromorphic algorithm development have shown that neural inspired architectures can efficiently solve scientific computing problems including graph decision problems and partial-integro differential equations (PIDEs). The latter requires the generation of a large number of samples from a stochastic process. While the Monte Carlo approximation of the solution of the PIDEs converges with an increasing number of sampled neuromorphic trajectories, the fidelity of samples from a given stochastic process using neuromorphic hardware requires verification. Such an exercise increases our trust in this emerging hardware and works toward unlocking its energy and scaling efficiency for scientific purposes such as synthetic data generation and stochastic simulation. In this paper, we focus our verification efforts on a one-dimensional Ornstein- Uhlenbeck stochastic differential equation. Using a discrete-time Markov chain approximation, we sample trajectories of the stochastic process across a variety of parameters on an Intel 8- Loihi chip Nahuku neuromorphic platform. Using relative entropy as a verification measure, we demonstrate that the random samples generated on Loihi are, in an average sense, acceptable. Finally, we demonstrate how Loihi's fidelity to the distribution changes as a function of the parameters of the Ornstein- Uhlenbeck equation, highlighting a trade-off between the lower-precision random number generation of the neuromorphic platform and our algorithm's ability to represent a discrete-time Markov chain.
Numerical simulations of Greenland and Antarctic ice sheets involve the solution of large-scale highly nonlinear systems of equations on complex shallow geometries. This work is concerned with the construction of Schwarz preconditioners for the solution of the associated tangent problems, which are challenging for solvers mainly because of the strong anisotropy of the meshes and wildly changing boundary conditions that can lead to poorly constrained problems on large portions of the domain. Here, two-level GDSW (Generalized Dryja–Smith–Widlund) type Schwarz preconditioners are applied to different land ice problems, i.e., a velocity problem, a temperature problem, as well as the coupling of the former two problems. We employ the MPI-parallel implementation of multi-level Schwarz preconditioners provided by the package FROSch (Fast and Robust Schwarz)from the Trilinos library. The strength of the proposed preconditioner is that it yields out-of-the-box scalable and robust preconditioners for the single physics problems. To our knowledge, this is the first time two-level Schwarz preconditioners are applied to the ice sheet problem and a scalable preconditioner has been used for the coupled problem. The pre-conditioner for the coupled problem differs from previous monolithic GDSW preconditioners in the sense that decoupled extension operators are used to compute the values in the interior of the sub-domains. Several approaches for improving the performance, such as reuse strategies and shared memory OpenMP parallelization, are explored as well. In our numerical study we target both uniform meshes of varying resolution for the Antarctic ice sheet as well as non uniform meshes for the Greenland ice sheet are considered. We present several weak and strong scaling studies confirming the robustness of the approach and the parallel scalability of the FROSch implementation. Among the highlights of the numerical results are a weak scaling study for up to 32 K processor cores (8 K MPI-ranks and 4 OpenMP threads) and 566 M degrees of freedom for the velocity problem as well as a strong scaling study for up to 4 K processor cores (and MPI-ranks) and 68 M degrees of freedom for the coupled problem.
This report presents the results of a collaborative effort under the Verification, Validation, and Uncertainty Quantification (VVUQ) thrust area of the North American Energy Resilience Model (NAERM) program. The goal of the effort described in this report was to integrate the Dakota software with the NAERM software framework to demonstrate sensitivity analysis of a co-simulation for NAERM.
Software development for high-performance scientific computing continues to evolve in response to increased parallelism and the advent of on-node accelerators, in particular GPUs. While these hardware advancements have the potential to significantly reduce turnaround times, they also present implementation and design challenges for engineering codes. We investigate the use of two strategies to mitigate these challenges: the Kokkos library for performance portability across disparate architectures, and the DARMA/vt library for asynchronous many-task scheduling. We investigate the application of Kokkos within the NimbleSM finite element code and the LAMÉ constitutive model library. We explore the performance of DARMA/vt applied to NimbleSM contact mechanics algorithms. Software engineering strategies are discussed, followed by performance analyses of relevant solid mechanics simulations which demonstrate the promise of Kokkos and DARMA/vt for accelerated engineering simulators.
We consider the integral definition of the fractional Laplacian and analyze a linearquadratic optimal control problem for the so-called fractional heat equation; control constraints are also considered. We derive existence and uniqueness results, first order optimality conditions, and regularity estimates for the optimal variables. To discretize the state equation we propose a fully discrete scheme that relies on an implicit finite difference discretization in time combined with a piecewise linear finite element discretization in space. We derive stability results and a novel L2(0, T;L2(Ω)) a priori error estimate. On the basis of the aforementioned solution technique, we propose a fully discrete scheme for our optimal control problem that discretizes the control variable with piecewise constant functions, and we derive a priori error estimates for it. We illustrate the theory with one- and two-dimensional numerical experiments.
In this work, we show that reduced communication algorithms for distributed stochastic gradient descent improve the time per epoch and strong scaling for the Generalized Canonical Polyadic (GCP) tensor decomposition, but with a cost, achieving convergence becomes more difficult. The implementation, based on MPI, shows that while one-sided algorithms offer a path to asynchronous execution, the performance benefits of optimized allreduce are difficult to best.
Gate set tomography (GST) is a protocol for detailed, predictive characterization of logic operations (gates) on quantum computing processors. Early versions of GST emerged around 2012-13, and since then it has been refined, demonstrated, and used in a large number of experiments. This paper presents the foundations of GST in comprehensive detail. The most important feature of GST, compared to older state and process tomography protocols, is that it is calibration-free. GST does not rely on pre-calibrated state preparations and measurements. Instead, it characterizes all the operations in a gate set simultaneously and self-consistently, relative to each other. Long sequence GST can estimate gates with very high precision and efficiency, achieving Heisenberg scaling in regimes of practical interest. In this paper, we cover GST’s intellectual history, the techniques and experiments used to achieve its intended purpose, data analysis, gauge freedom and fixing, error bars, and the interpretation of gauge-fixed estimates of gate sets. Our focus is fundamental mathematical aspects of GST, rather than implementation details, but we touch on some of the foundational algorithmic tricks used in the pyGSTi implementation.
Environmental Barrier Coatings (EBC) protect ceramic matrix composites from exposure to high temperature moisture present in turbine operation through their dense top coats. However, moisture is able to diffuse and oxidize the Si bond coat to form the Thermally Grown Oxide (TGO), a layer of SiO2 where the incorporation of O causes swelling and stress. At sufficient TGO-based swelling, the EBC will fail due to increased damage such as delamination. A multiscale simulation framework has been developed to link operating conditions of a high-performance turbine to the failure modes of the EBC. Computational fluid dynamics (CFD) simulations of the E3 turbine were performed and compared to prior literature data to demonstrate the fidelity of the Loci/CHEM software to determine the flow conditions on the turbine blade surface. Boundary condition data of pressure and heat flux were then determined with the CFD simulations, providing the temperature at the bond coat. Peridynamics was used to model the microscale TGO growth. A swelling model that links moisture concentration to strain at the TGO due to the volume increase from oxidation was demonstrated, coupling moisture transport to localized strain and directly observing TGO growth and the corresponding damage. This framework is generalized and can be adapted to a range of EBC microstructures and operating conditions.
Poisson Tensor Factorization (PTF) is an important data analysis method for analyzing patterns and relationships in multiway count data. In this work, we consider several algorithms for computing a low-rank PTF of tensors with sparse count data values via maximum likelihood estimation. Such an approach reduces to solving a nonlinear, non-convex optimization problem, which can leverage considerable parallel computation due to the structure of the problem. However, since the maximum likelihood estimator corresponds to the global minimizer of this optimization problem, it is important to consider how effective methods are at both leveraging this inherent parallelism as well as computing a good approximation to the global minimizer. In this work we present comparisons of multiple methods for PTF that illustrate the tradeoffs in computational efficiency and accurately computing the maximum likelihood estimator. We present results using synthetic and real-world data tensors to demonstrate some of the challenges when choosing a method for a given tensor.
In this work we evaluated the effects that equations of state and strength models have on SCJ development using the Sandia National Laboratories multiphysics shock code, ALEGRA. Results were quantified using a Lagrangian tracer particle following liner collapse, passing through the compression zone, and flowing into the jet tip. We found consistent results among several EOS: 3320, 3331, and 3337. The 3325 EOS generated a measurable low density and hollow region near the jet tip which appears to be reflected in a lower internal energy of the jet. At this time, we cannot tell, experimentally, if such a hollow region exists. The 3337 EOS is recent, well documented [6], and produces results similar to 3320 [3]. The various strength models produced more noticeable differences. In terms of internal energy and temperature, SGL had the largest values followed by PTW, ZA, and finally JC and MTS, which were quite similar to each other. We looked at melt conditions in the SGL and JC models using the 3337 EOS. The SGL model reported a liquid region along the jet axis all the way to the tip-seemingly consistent with experiment-while the JC model does not indicate any phase transition. None of the other yield models indicated melt along the jet axis. For all EOS and strength models, we found similar results for the velocity history of the jet tip as measured against experiment using photon Dopper velocimetry.
Boolean functions and binary arithmetic operations are central to standard computing paradigms. Accordingly, many advances in computing have focused upon how to make these operations more efficient as well as exploring what they can compute. To best leverage the advantages of novel computing paradigms it is important to consider what unique computing approaches they offer. However, for any special-purpose co-processor, Boolean functions and binary arithmetic operations are useful for, among other things, avoiding unnecessary I/O on-and-off the co-processor by pre- and post-processing data on-device. This is especially true for spiking neuromorphic architectures where these basic operations are not fundamental low-level operations. Instead, these functions require specific implementation. Here we discuss the implications of an advantageous streaming binary encoding method as well as a handful of circuits designed to exactly compute elementary Boolean and binary operations.
We have extended the computational singular perturbation (CSP) method to differential algebraic equation (DAE) systems and demonstrated its application in a heterogeneous-catalysis problem. The extended method obtains the CSP basis vectors for DAEs from a reduced Jacobian matrix that takes the algebraic constraints into account. We use a canonical problem in heterogeneous catalysis, the transient continuous stirred tank reactor (T-CSTR), for illustration. The T-CSTR problem is modelled fundamentally as an ordinary differential equation (ODE) system, but it can be transformed to a DAE system if one approximates typically fast surface processes using algebraic constraints for the surface species. We demonstrate the application of CSP analysis for both ODE and DAE constructions of a T-CSTR problem, illustrating the dynamical response of the system in each case. We also highlight the utility of the analysis in commenting on the quality of any particular DAE approximation built using the quasi-steady state approximation (QSSA), relative to the ODE reference case.
Aeroengines ingest foreign object debris such as sand, which eventually erode components through repeated impacts. Due to the wide feature space, modeling and simulations are needed to rapidly assess the erosion behavior of materials such as composites. Peridynamic simulations were performed to analyze erosion of SiC/SiC composite due to sand impacts, which gives direct insight into the impact erosion mechanism and amounts. The erosion data was strongly correlated to impact velocity and angle, providing predictive equations.
Bayesian optimization (BO) is a flexible and powerful framework that is suitable for computationally expensive simulation-based applications and guarantees statistical convergence to the global optimum. While remaining as one of the most popular optimization methods, its capability is hindered by the size of data, the dimensionality of the considered problem, and the nature of sequential optimization. These scalability issues are intertwined with each other and must be tackled simultaneously. In this work, we propose the Scalable3-BO framework, which employs sparse GP as the underlying surrogate model to scope with Big Data and is equipped with a random embedding to efficiently optimize high-dimensional problems with low effective dimensionality. The Scalable3-BO framework is further leveraged with asynchronous parallelization feature, which fully exploits the computational resource on HPC within a computational budget. As a result, the proposed Scalable3-BO framework is scalable in three independent perspectives: with respect to data size, dimensionality, and computational resource on HPC. The goal of this work is to push the frontiers of BO beyond its well-known scalability issues and minimize the wall-clock waiting time for optimizing high-dimensional computationally expensive applications. We demonstrate the capability of Scalable3-BO with 1 million data points, 10,000-dimensional problems, with 20 concurrent workers in an HPC environment.
We propose a learning algorithm for discovering unknown parameterized dynamical systems by using observational data of the state variables. Our method is built upon and extends the recent work of discovering unknown dynamical systems, in particular those using a deep neural network (DNN). We propose a DNN structure, largely based upon the residual network (ResNet), to not only learn the unknown form of the governing equation but also to take into account the random effect embedded in the system, which is generated by the random parameters. Once the DNN model is successfully constructed, it is able to produce system prediction over a longer term and for arbitrary parameter values. For uncertainty quantification, it allows us to conduct uncertainty analysis by evaluating solution statistics over the parameter space.
This paper develops a novel limited-memory method to solve dynamic optimization problems. The memory requirements for such problems often present a major obstacle, particularly for problems with PDE constraints such as optimal flow control, full waveform inversion, and optical tomography. In these problems, PDE constraints uniquely determine the state of a physical system for a given control; the goal is to find the value of the control that minimizes an objective. While the control is often low dimensional, the state is typically more expensive to store. This paper suggests using randomized matrix approximation to compress the state as it is generated and shows how to use the compressed state to reliably solve the original dynamic optimization problem. Concretely, the compressed state is used to compute approximate gradients and to apply the Hessian to vectors. The approximation error in these quantities is controlled by the target rank of the sketch. This approximate first- and second-order information can readily be used in any optimization algorithm. As an example, we develop a sketched trust-region method that adaptively chooses the target rank using a posteriori error information and provably converges to a stationary point of the original problem. Numerical experiments with the sketched trust-region method show promising performance on challenging problems such as the optimal control of an advection-reaction-diffusion equation and the optimal control of fluid flow past a cylinder.
In this paper, we introduce and analyze a new class of optimal control problems constrained by elliptic equations with uncertain fractional exponents. We utilize risk measures to formulate the resulting optimization problem. We develop a functional analytic framework, study the existence of solution, and rigorously derive the first-order optimality conditions. Additionally, we employ a sample-based approximation for the uncertain exponent and the finite element method to discretize in space. We prove the rate of convergence for the optimal risk neutral controls when using quadrature approximation for the uncertain exponent and conclude with illustrative examples.
Process-structure linkage is one of the most important topics in materials science due to the fact that virtually all information related to the materials, including manufacturing processes, lies in the microstructure itself. Therefore, to learn more about the process, one must start by thoroughly examining the microstructure. This gives rise to inverse problems in the context of process-structure linkages, which attempt to identify the processes that were used to manufacturing the given microstructure. In this work, we present an inverse problem for structure-process linkages which we solve using asynchronous parallel Bayesian optimization which exploits parallel computing resources. We demonstrate the effectiveness of the method using kinetic Monte Carlo model for grain growth simulation.
Process-structure-property relationships are the hallmark of materials science. Many integrated computational materials engineering (ICME) models have been developed at multiple length-scales and time-scales, where uncertainty quantification (UQ) plays an important role in quality assurance. In this paper, we applied our previous work [39] to learn a distribution of microstructure features that are consistent in the sense that the forward propagation of this distribution through a crystal plasticity finite element model (CPFEM) matches a target distribution on materials properties, which is given beforehand. To demonstrate the approach, DAMASK and DREAM.3D are employed to construct Hall-Petch relationship for a twinning-induced plasticity (TWIP) steel, where the average grain size distribution is inferred, given a distribution of offset yield strength.
This work proposes an approach for latent-dynamics learning that exactly enforces physical conservation laws. The method comprises two steps. First, the method computes a low-dimensional embedding of the high-dimensional dynamical-system state using deep convolutional autoencoders. This defines a low-dimensional nonlinear manifold on which the state is subsequently enforced to evolve. Second, the method defines a latent-dynamics model that associates with the solution to a constrained optimization problem. Here, the objective function is defined as the sum of squares of conservation-law violations over control volumes within a finite-volume discretization of the problem; nonlinear equality constraints explicitly enforce conservation over prescribed subdomains of the problem. Under modest conditions, the resulting dynamics model guarantees that the time-evolution of the latent state exactly satisfies conservation laws over the prescribed subdomains.
Solving dense systems of linear equations is essential in applications encountered in physics, mathematics, and engineering. This paper describes our current efforts toward the development of the ADELUS package for current and next generation distributed, accelerator-based, high-performance computing platforms. The package solves dense linear systems using partial pivoting LU factorization on distributed-memory systems with CPUs/GPUs. The matrix is block-mapped onto distributed memory on CPUs/GPUs and is solved as if it was torus-wrapped for an optimal balance of computation and communication. A permutation operation is performed to restore the results so the torus-wrap distribution is transparent to the user. This package targets performance portability by leveraging the abstractions provided in the Kokkos and Kokkos Kernels libraries. Comparison of the performance gains versus the state-of-the-art SLATE and DPLASMA GESV functionalities on the Summit supercomputer are provided. Preliminary performance results from large-scale electromagnetic simulations using ADELUS are also presented. The solver achieves 7.7 Petaflops on 7600 GPUs of the Sierra supercomputer translating to 16.9% efficiency.
The purpose of this paper is to study a Helmholtz problem with a spectral fractional Laplacian, instead of the standard Laplacian. Recently, it has been established that such a fractional Helmholtz problem better captures the underlying behavior in geophysical electromagnetics. We establish the well-posedness and regularity of this problem. We introduce a hybrid spectral-finite element approach to discretize it and show well-posedness of the discrete system. In addition, we derive a priori discretization error estimates. Finally, we introduce an efficient solver that scales as well as the best possible solver for the classical integer-order Helmholtz equation. We conclude with several illustrative examples that confirm our theoretical findings.
Performance variation diagnosis in High-Performance Computing (HPC) systems is a challenging problem due to the size and complexity of the systems. Application performance variation leads to premature termination of jobs, decreased energy efficiency, or wasted computing resources. Manual root-cause analysis of performance variation based on system telemetry has become an increasingly time-intensive process as it relies on human experts and the size of telemetry data has grown. Recent methods use supervised machine learning models to automatically diagnose previously encountered performance anomalies in compute nodes. However, supervised machine learning models require large labeled data sets for training. This labeled data requirement is restrictive for many real-world application domains, including HPC systems, because collecting labeled data is challenging and time-consuming, especially considering anomalies that sparsely occur. This paper proposes a novel semi-supervised framework that diagnoses previously encountered performance anomalies in HPC systems using a limited number of labeled data points, which is more suitable for production system deployment. Our framework first learns performance anomalies’ characteristics by using historical telemetry data in an unsupervised fashion. In the following process, we leverage supervised classifiers to identify anomaly types. While most semi-supervised approaches do not typically use anomalous samples, our framework takes advantage of a few labeled anomalous samples to classify anomaly types. We evaluate our framework on a production HPC system and on a testbed HPC cluster. We show that our proposed framework achieves 60% F1-score on average, outperforming state-of-the-art supervised methods by 11%, and maintains an average 0.06% anomaly miss rate.
Reproducing kernel (RK) approximations are meshfree methods that construct shape functions from sets of scattered data. We present an asymptotically compatible (AC) RK collocation method for nonlocal diffusion models with Dirichlet boundary condition. The numerical scheme is shown to be convergent to both nonlocal diffusion and its corresponding local limit as nonlocal interaction vanishes. The analysis is carried out on a special family of rectilinear Cartesian grids for a linear RK method with designed kernel support. The key idea for the stability of the RK collocation scheme is to compare the collocation scheme with the standard Galerkin scheme, which is stable. In addition, assembling the stiffness matrix of the nonlocal problem requires costly computational resources because high-order Gaussian quadrature is necessary to evaluate the integral. We thus provide a remedy to the problem by introducing a quasi-discrete nonlocal diffusion operator for which no numerical quadrature is further needed after applying the RK collocation scheme. The quasi-discrete nonlocal diffusion operator combined with RK collocation is shown to be convergent to the correct local diffusion problem by taking the limits of nonlocal interaction and spatial resolution simultaneously. The theoretical results are then validated with numerical experiments. We additionally illustrate a connection between the proposed technique and an existing optimization based approach based on generalized moving least squares.
Neural architecture search (NAS) has emerged as an algorithmic method of developing neural network architectures. Weight Agnostic Neural Networks (WANNs) are an evolutionary-based NAS approach. Fundamentally, WANNs find network structures that are relatively insensitive to shifts in weight values and are typically much smaller than an equivalent performance dense network. Here, we extend the WANN framework to search for spiking circuits and in doing so investigate whether these circuit motifs can also yield task performance that is weight agnostic. We analyze properties such as the complexity of the solution, as well as performance. Our results successfully show the performance of spiking WANNs on several exemplar tasks.
Both the data science and scientific computing communities are embracing GPU acceleration for their most demanding workloads. For scientific computing applications, the massive volume of code and diversity of hardware platforms at supercomputing centers has motivated a strong effort toward performance portability. This property of a program, denoting its ability to perform well on multiple architectures and varied datasets, is heavily dependent on the choice of parallel programming model and which features of the programming model are used. In this paper, we evaluate performance portability in the context of a data science workload in contrast to a scientific computing workload, evaluating the same sparse matrix kernel on both. Among our implementations of the kernel in different performance-portable programming models, we find that many struggle to consistently achieve performance improvements using the GPU compared to simple one-line OpenMP parallelization on high-end multicore CPUs. We show one that does, and its performance approaches and sometimes even matches that of vendor-provided GPU math libraries.
Polynomial preconditioning can improve the convergence of the Arnoldi method for computing eigenvalues. Such preconditioning significantly reduces the cost of orthogonalization; for difficult problems, it can also reduce the number of matrix-vector products. Parallel computations can particularly benefit from the reduction of communication-intensive operations. The GMRES algorithm provides a simple and effective way of generating the preconditioning polynomial. For some problems high degree polynomials are especially effective, but they can lead to stability problems that must be mitigated. A two-level "double polynomial preconditioning"strategy provides an effective way to generate high-degree preconditioners.
Determining process–structure–property linkages is one of the key objectives in material science, and uncertainty quantification plays a critical role in understanding both process–structure and structure–property linkages. In this work, we seek to learn a distribution of microstructure parameters that are consistent in the sense that the forward propagation of this distribution through a crystal plasticity finite element model matches a target distribution on materials properties. This stochastic inversion formulation infers a distribution of acceptable/consistent microstructures, as opposed to a deterministic solution, which expands the range of feasible designs in a probabilistic manner. To solve this stochastic inverse problem, we employ a recently developed uncertainty quantification framework based on push-forward probability measures, which combines techniques from measure theory and Bayes’ rule to define a unique and numerically stable solution. This approach requires making an initial prediction using an initial guess for the distribution on model inputs and solving a stochastic forward problem. To reduce the computational burden in solving both stochastic forward and stochastic inverse problems, we combine this approach with a machine learning Bayesian regression model based on Gaussian processes and demonstrate the proposed methodology on two representative case studies in structure–property linkages.