Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “compute intensive problems”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

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

At least 181 records · Page 10

2009 ALCF annual report.

This year the Argonne Leadership Computing Facility (ALCF) delivered nearly 900 million core hours of science. The research conducted at their leadership class facility touched our lives in both minute and massive ways - whether it was studying the catalytic properties of gold nanoparticles, predicting protein structures, or unearthing the secrets of exploding stars. The authors remained true to their vision to act as the forefront computational center in extending science frontiers by solving pressing problems for our nation. Our success in this endeavor was due mainly to the Department of Energy's (DOE) INCITE (Innovative and Novel Computational Impact on Theory and Experiment) program. The program awards significant amounts of computing time to computationally intensive, unclassified research projects that can make high-impact scientific advances. This year, DOE allocated 400 million hours of time to 28 research projects at the ALCF. Scientists from around the world conducted the research, representing such esteemed institutions as the Princeton Plasma Physics Laboratory, National Institute of Standards and Technology, and European Center for Research and Advanced Training in Scientific Computation. Argonne also provided Director's Discretionary allocations for research challenges, addressing such issues as reducing aerodynamic noise, critical for next-generation 'green' energy systems. Intrepid - the ALCF's 557-teraflops IBM Blue/Gene P supercomputer - enabled astounding scientific solutions and discoveries. Intrepid went into full production five months ahead of schedule. As a result, the ALCF nearly doubled the days of production computing available to the DOE Office of Science, INCITE awardees, and Argonne projects. One of the fastest supercomputers in the world for open science, the energy-efficient system uses about one-third as much electricity as a machine of comparable size built with more conventional parts. In October 2009, President Barack Obama recognized the excellence of the entire Blue Gene series by awarding it to the National Medal of Technology and Innovation. Other noteworthy achievements included the ALCF's collaboration with the National Energy Research Scientific Computing Center (NERSC) to examine cloud computing as a potential new computing paradigm for scientists. Named Magellan, the DOE-funded initiative will explore which science application programming models work well within the cloud, as well as evaluate the challenges that come with this new paradigm. The ALCF obtained approval for its next-generation machine, a 10-petaflops system to be delivered in 2012. This system will allow us to resolve ever more pressing problems, even more expeditiously through breakthrough science in the years to come.

59 BASIC BIOLOGICAL SCIENCES↗

Energy Efficient Stochastic-Based Deep Spiking Neural Networks for Sparse Datasets

With large deep neural networks (DNNs) neces- sary to solve complex and data-intensive problems, energy efficiency is a key bottleneck for effectively deploying DL in the real world. Deep spiking NNs have gained much research at- tention recently due to the interest in building biological neural networks and the availability of neuromorphic platforms, which can be orders of magnitude more energy efficient compared to CPUs and GPUs. Although spiking NNs have proven to be an efficient technique for solving many machine learning and computer vision problems, to the best of our knowledge, this is the first attempt to adapt spiking NNs to sparse datasets. In this paper, we study the behaviour of spiking NNs in handling NLP datasets and the sparsity in their data representation. Then, we propose a novel framework for spiking NN using the concept of stochastic computing. Specifically, instead of generating spike trains with firing rates proportional to the intensity of each value in the feature set separately, the whole feature set is treated as a distribution function and a stochastic spiking train that follow this distribution is generated. This framework reduces the connectivity between NN layers from O(N) to O(logN). Also, it encodes input data differently and make suitable to handle sparse datasets. Finally, the framework achieves high energy efficiency since it uses Integrate and Fire neurons same as conventional spiking NNs. The results show that our proposed stochastic-based SNN achieves nearly the same accuracy as the original DNN on MNIST dataset, and it has better performance than state- of-the-art SNN. Besides that stochastic-based SNN is energy efficient, where the fully connected DNN, the conventional SNN, and the data normalized SNN consume 38.24, 1.83, and 1.85 -times more energy than the stochastic-based SNN, respectively. For sparse datasets, including IMDb and In-House clinical datasets, stochastic-based SNN achieves performance comparable to that of the conventional DNN. However, the conventional spiking NN has a significant decline in classification accuracy.

Alawad, Mohammed↗

Design of detectors at the electron ion collider with artificial intelligence

Abstract Artificial Intelligence (AI) for design is a relatively new but active area of research across many disciplines. Surprisingly when it comes to designing detectors with AI this is an area at its infancy. The electron ion collider is the ultimate machine to study the strong force. The EIC is a large-scale experiment with an integrated detector that extends for about ±35 meters to include the central, far-forward, and far-backward regions. The design of the central detector is made by multiple sub-detectors, each in principle characterized by a multidimensional design space and multiple design criteria also called objectives. Simulations with Geant4 are typically compute intensive, and the optimization of the detector design may include non-differentiable terms as well as noisy objectives. In this context, AI can offer state of the art solutions to solve complex combinatorial problems in an efficient way. In particular, one of the proto-collaborations, ECCE, has explored during the detector proposal the possibility of using multi-objective optimization to design the tracking system of the EIC detector. This document provides an overview of these techniques and recent progress made during the EIC detector proposal. Future high energy nuclear physics experiments can leverage AI-based strategies to design more efficient detectors by optimizing their performance driven by physics criteria and minimizing costs for their realization.

Instruments & Instrumentation↗

Data and Thread Placement in NUMA Architectures: A Statistical Learning Approach

Nowadays, NUMA architectures are common in compute-intensive systems. Achieving high performance for multi-threaded application requires both a careful placement of threads on computing units and a thorough allocation of data in memory. Finding such a placement is a hard problem to solve, because performance depends on complex interactions in several layers of the memory hierarchy. In this paper we propose a black-box approach to decide if an application execution time can be impacted by the placement of its threads and data, and in such a case, to choose the best placement strategy to adopt. We show that it is possible to reach near-optimal placement policy selection. Furthermore, solutions work across several recent processor architectures and decisions can be taken with a single run of low overhead profiling.

data↗

A survey on processing-in-memory techniques: Advances and challenges

Processing-in-memory (PIM) techniques have gained much attention from computer architecture researchers, and significant research effort has been invested in exploring and developing such techniques. Increasing the research activity dedicated to improving PIM techniques will hopefully help deliver PIM’s promise to solve or significantly reduce memory access bottleneck problems for memory-intensive applications. We also believe it is imperative to track the advances made in PIM research to identify open challenges and enable the research community to make informed decisions and adjust future research directions. In this survey, we analyze recent studies that explored PIM techniques, summarize the advances made, compare recent PIM architectures, and identify target application domains and suitable memory technologies. We also discuss proposals that address unresolved issues of PIM designs (e.g., address translation/mapping of operands, workload analysis to identify application segments that can be accelerated with PIM, OS/runtime support, and coherency issues that must be resolved to incorporate PIM). We believe this work can serve as a useful reference for researchers exploring PIM techniques.

97 MATHEMATICS AND COMPUTING↗

Geant4 Computing Performance Benchmarking and Monitoring

Performance evaluation and analysis of large scale computing applications is essential for optimal use of resources. As detector simulation is one of the most compute intensive tasks and Geant4 is the simulation toolkit most widely used in contemporary high energy physics (HEP) experiments, it is important to monitor Geant4 through its development cycle for changes in computing performance and to identify problems and opportunities for code improvements. All Geant4 development and public releases are being profiled with a set of applications that utilize different input event samples, physics parameters, and detector configurations. Results from multiple benchmarking runs are compared to previous public and development reference releases to monitor CPU and memory usage. Observed changes are evaluated and correlated with code modifications. Besides the full summary of call stack and memory footprint, a detailed call graph analysis is available to Geant4 developers for further analysis. The set of software tools used in the performance evaluation procedure, both in sequential and multi-threaded modes, include FAST, IgProf and Open|Speedshop. In conclusion, the scalability of the CPU time and memory performance in multi-threaded application is evaluated by measuring event throughput and memory gain as a function of the number of threads for selected event samples.

97 MATHEMATICS AND COMPUTING↗

SUNDIALS Multiphysics+MPIManyVector Performance Testing

In this report we document performance test results on a SUNDIALS-based multiphysics demonstration application. We aim to assess the large-scale parallel performance of new capabilities that have been added to the SUNDIALS suite of time integrators and nonlinear solvers in recent years under funding from both the Exascale Computing Project (ECP) and the Scientific Discovery through Advanced Scientific (SciDAC) program, specifically: 1. SUNDIALS’ new MPIManyVector module, that allows extreme flexibility in how a solution "vector" is staged on computational resources. 2. ARKode’s new multirate integration module, MRIStep, allowing high-order accurate calculations that subcycle "fast" processes within "slow" ones. 3. SUNDIALS’ new flexible linear solver interfaces, that allow streamlined specification of problem-specific linear solvers. 4. SUNDIALS’ new N_Vector additions of "fused" vector operations (to increase arithmetic intensity) and separation of reduction operations into "local" and "global" versions (to reduce latency by combining multiple reductions into a single MPI_Allreduce call). We anticipate that subsequent reports will extend this work to investigate a variety of other new features, including SUNDIALS’ generic SUNNonlinearSolver interface and accelerator-enabled N Vector modules, and upcoming MRIStep extensions to support custom "fast" integrators (that leverage problem structure) and IMEX integration of the "slow" time scale (to add diffusion).

97 MATHEMATICS AND COMPUTING↗

A Scalable Parallel Wideband MLFMA for Efficient Electromagnetic Simulations on Large Scale Clusters

The development of the multilevel fast multipole algorithm (MLFMA) and its multiscale variants have enabled the use of integral equation (IE) based solvers to compute scattering from complicated structures. Development of scalable parallel algorithms, to extend the reach of these solvers, has been a topic of intense research for about a decade. In this paper, we present a new algorithm for parallel implementation of IE solver that is augmented with a wideband MLFMA and scalable on large number of processors. The wideband MLFMA employed here, to handle multiscale problems, is a hybrid combination of the accelerated Cartesian expansion (ACE) and the classical MLFMA. The salient feature of the presented parallel algorithm is that it is implicitly load balanced and exhibits higher performance. This is achieved by developing a strategy to partition the MLFMA tree, and hence the associated computations, in a self-similar fashion among the parallel processors. As detailed in the paper, the algorithm employs both spatial and direction partitioning approaches in a flexible manner to ensure scalable performance. Plethora of results are presented here to exhibit the scalability of this algorithm on 512 and more processors.

99 GENERAL AND MISCELLANEOUS↗

Machine learning for domain transfer between simulated and experimental 2D X-ray diffraction patterns using generative adversarial networks

X-ray diffraction (XRD) is a well-established technique for analyzing materials at an atomic level. Dynamic compression experiments (DCE), in which materials are subject to extreme pressures, can provide fundamental understanding to pressure-induced phase transitions and compression of the crystal lattice. The analysis of XRD patterns from highly compressed samples is non-trivial given the sparsity of data, high experimental costs, and the fact that the data is often marred with X-ray background and other artifacts. While accurate computational frameworks exist, they solve the forward problem—from structures and orientations to XRD patterns. Solving the inverse problem for 2D experimental diffraction patterns is currently a complex manual process of matching and comparing experimentally observed patterns to computationally generated ones. Machine learning is a promising tool for automating the matching process but often requires data-intensive architectures. Here, in this study, we use a CycleGAN to translate the domain of limited experimental data to a domain in which there is readily available simulated data. This domain shift allows data-intensive machine learning models that have only been trained on simulated XRD patterns to be used in the analysis of experiments.

Brozak, Samantha Jean [Sandia National Laboratorie↗

Evaluation of low-exergy heating and cooling systems and topology optimization for deep energy savings at the urban district level

District energy systems have the potential to achieve deep energy savings by leveraging the density and diversity of loads in urban districts. However, planning and adoption of district thermal energy systems is hindered by the analytical burden and high infrastructure costs. It is hypothesized that network topology optimization would enable wider adoption of advanced (ambient temperature) district thermal energy systems, resulting in energy savings. In this study, energy modeling is used to compare the energy performance of “conventional” and “advanced” district thermal energy systems at the urban district level, and a partial exhaustive search is used to evaluate a heuristic for the topology optimization problem. For the prototypical district considered, advanced district thermal energy systems mated with low-exergy building heating and cooling systems achieved a source energy use intensity that was 49% lower than that of conventional systems. The minimal spanning tree heuristic was demonstrated to be effective for the network topology optimization problem in the context of a prototypical district, and contributes to mitigating the problem’s computational complexity. The work presented in this paper demonstrates the potential of advanced district thermal energy systems to achieve deep energy savings, and advances to addressing barriers to their adoption through topology optimization.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Analysis of Mode III Elastodynamic Cracked Plane using the Fractal Two-Level Finite Element Method

In this study, the fractal two-level finite element method, which has mainly been used for static cracked plane problems, is applied to the cracked plane problem. Using the transformation process in the proposed method, the infinite dimension of the finite element matrices that are assembled for a singular region is made finite in terms of the dynamics stress intensity factors directly, and thus the computational time can be reduced significantly. The Newmark time integration scheme is then used to obtain the dynamic stress intensity factors. The results from the proposed method are in reasonable agreement with those of classical methods. The main drawback of the time integration scheme is that numerical oscillations are induced in some cases.

97 MATHEMATICS AND COMPUTING↗

Tall Tower Wind Energy Monitoring and Numerical Model Validation in Northern Nevada

The main objectives of this project were to conduct a tall-tower and sodar field campaign in complex terrain, investigate wind properties relevant to wind energy assessment, and evaluate high-resolution models with fixed and adaptive grid structures. Two 60-m towers at Virginia Peak ridges near Washoe Valley, Nevada, were instrumented with cup and vane anemometers as well as sonic anemometers, and an acoustic sounder (hereafter sodar) was installed near one of the towers. The towers were located 2,700 m apart with a vertical distance of 140 m elevation between their bases. Each tower had a downhill exposure of rolling complex terrain, with the nearby valley floor 3,200 m to the west and 800 m below the summit. Cup anemometers were installed at both towers at 20, 40, and 60 m, wind vanes at 20 and 60 m, and sonic anemometers at 20 and 60 m. The sodar measurements were nominally provided every 10 m in vertical distance from 40 to 200 m with the quality of the data generally decreasing with height. Surface air temperature, atmospheric pressure, and radiation measurements were conducted at 1.5 m AGL at both of the towers. Although the plan was to conduct a 1-year period of data collection, we extended the period (October 5, 2012 through February 24, 2014) to cover for possible data loss from instrument or communication problems. We also present a preliminary analysis of the towers and sodar data, including a detailed inventory of available and missing data as well as outliers. The analysis additionally includes calculation of the Weibull parameters, turbulence intensity, and initial computation of wind power density at various heights.

17 WIND ENERGY↗

The reduced basis method for the electric field integral equation

We introduce the reduced basis method (RBM) as an efficient tool for parametrized scattering problems in computational electromagnetics for problems where field solutions are computed using a standard Boundary Element Method (BEM) for the parametrized electric field integral equation (EFIE). This combination enables an algorithmic cooperation which results in a two step procedure. The first step consists of a computationally intense assembling of the reduced basis, that needs to be effected only once. In the second step, we compute output functionals of the solution, such as the Radar Cross Section (RCS), independently of the dimension of the discretization space, for many different parameter values in a many-query context at very little cost. Parameters include the wavenumber, the angle of the incident plane wave and its polarization.

97 MATHEMATICS AND COMPUTING↗

Rapid 3D nanoscale coherent imaging via physics-aware deep learning

Phase retrieval, the problem of recovering lost phase information from measured intensity alone, is an inverse problem that is widely faced in various imaging modalities ranging from astronomy to nanoscale imaging. The current process of phase recovery is iterative in nature. As a result, the image formation is time consuming and computationally expensive, precluding real-time imaging. Here, we use 3D nanoscale X-ray imaging as a representative example to develop a deep learning model to address this phase retrieval problem. We introduce 3D-CDI-NN, a deep convolutional neural network and differential programing framework trained to predict 3D structure and strain, solely from input 3D X-ray coherent scattering data. Our networks are designed to be “physics-aware” in multiple aspects; in that the physics of the X-ray scattering process is explicitly enforced in the training of the network, and the training data are drawn from atomistic simulations that are representative of the physics of the material. We further refine the neural network prediction through a physics-based optimization procedure to enable maximum accuracy at lowest computational cost. 3D-CDI-NN can invert a 3D coherent diffraction pattern to real-space structure and strain hundreds of times faster than traditional iterative phase retrieval methods. Our integrated machine learning and differential programing solution to the phase retrieval problem is broadly applicable across inverse problems in other application areas.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

Highly-scalable, physics-informed GANs for learning solutions of stochastic PDEs

Uncertainty quantification for forward and inverse problems is a central challenge across physical and biomedical disciplines. We address this challenge for the problem of modeling subsurface flow at the Hanford Site by combining stochastic computational models with observational data using physics-informed GAN models. The geographic extent, spatial heterogeneity, and multiple correlation length scales of the Hanford Site require training a computationally intensive GAN model to thousands of dimensions. We develop a highly optimized implementation that scales to {\herogpucount} NVIDIA Volta GPUs. We develop a hierarchical scheme based on a multi-player game-theoretic approach for exploiting domain parallelism, map discriminators and generators to multiple GPUs, and employ efficient communication schemes to ensure training stability and convergence. Our implementation scales to {\heronodecount} nodes on the Summit supercomputer with a \weakscalingeff\% scaling efficiency, achieving peak and sustained half-precision rates of {\heropeakperf} PF/s and {\herosustainedperf} PF/s.

physics informed ML, machine learning, generative ↗

Surrogate construction via weight parameterization of residual neural networks

Surrogate model development is a critical step for uncertainty quantification or other sample-intensive tasks for complex computational models. Here, in this work, we develop a multi-output surrogate form using a class of neural networks (NNs) that employ shortcut connections, namely Residual NNs (ResNets). ResNets are known to regularize the surrogate learning problem and improve the efficiency and accuracy of the resulting surrogate. Inspired by the continuous, Neural ODE analogy, we augment ResNets with weight parameterization strategy with respect to ResNet depth. Weight-parameterized ResNets regularize the NN surrogate learning problem and allow better generalization with a drastically reduced number of learnable parameters. We demonstrate that weight-parameterized ResNets are more accurate and efficient than conventional feed-forward multi-layer perceptron networks. We also compare various options for parameterization of the weights as functions of ResNet depth. We demonstrate the results on both synthetic examples and a large scale earth system model of interest.

97 MATHEMATICS AND COMPUTING↗

An iterative phase-space explicit discontinuous Galerkin method for stellar radiative transfer in extended atmospheres

In this work, a phase-space discontinuous Galerkin (PSDG) method is presented for the solution of stellar radiative transfer problems. It allows for greater adaptivity than competing methods without sacrificing generality. The method is extensively tested on a spherically symmetric, static, inverse-power-law scattering atmosphere. Results for different sizes of atmospheres and intensities of scattering agreed with asymptotic values. The exponentially decaying behavior of the radiative field in the diffusive-transparent transition region, and the forward peaking behavior at the surface of extended atmospheres were accurately captured. The integrodifferential equation of radiation transfer is solved iteratively by alternating between the radiative pressure equation and the original equation with the integral term treated as an energy density source term. In each iteration, the equations are solved via an explicit, flux-conserving, discontinuous Galerkin method. Finite elements are ordered in wave fronts perpendicular to the characteristic curves so that elemental linear algebraic systems are solved quickly by sweeping the phase space element by element. Two implementations of a diffusive boundary condition at the origin are demonstrated wherein the finite discontinuity in the radiation intensity is accurately captured by the proposed method. This allows for a consistent mechanism to preserve photon luminosity. The method was proved to be robust and fast, and a case is made for the adequacy of parallel processing. In addition to classical two-dimensional plots, results of normalized radiation intensity were mapped onto a log-polar surface exhibiting all distinguishing features of the problem studied.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Further Investigation of Employing Binomial Small-Sample Estimator in M-LROM

Reduced Order Modeling (ROM) is essential to enable the identification of the so-called active parameter subspace for high-fidelity models that are employed in computationally intensive analyses such as uncertainty characterization and data assimilation. Given the complexity of models for real-world application, brute force methods to identify the active subspace may not be computationally efficient to realize the value of reduction, since in realistic problems, the full scale model must be executed few hundred times, which may be computationally taxing. Earlier work has introduced a new approach, which represents a further development of our previous work on ROM, referred to as multi-level ROM (M-LROM). In M-LROM, approximate models are employed to identify the active subspace instead of the full scale model. In this summary, we present our recent results on M-LROM, specifically in regard to the construction of reliable error metrics that can bound the errors resulting from the reduction. The idea is to employ a lower fidelity model, i.e., an approximation of the full scale or high fidelity model, to build the active subspace, and via the proposed error bound, the quality of the active subspace can be determined by calculating an upper bound on the errors resulting from the reduction. For this approach to be practical, the error bound must be realistic, implying that it should be close to the actual errors expected, and not much higher. Therefore, an important goal of this summary is to compare different error bounds in terms of the relation between their estimated bounds to the actual errors. In particular, we compare the proposed error bound which numerically construct an estimator using binomial distribution to the widely used small-sample norm estimators. Unlike most probabilistic error estimation methods, the proposed bound employs a binomial distribution to construct the estimator whereas in other conventional methods, the distribution employed is typically the standard Gaussian, normalized over a unit hypersphere as initially proposed by the seminal paper of J. Dixon on random matrix theory. Previous publications provided evidence that the binomial distribution calculates more realistic bounds as compared to the Gaussian distribution, which tends to be conservative, implying that the bound is much bigger than the actual errors, which diminishes the value of the reduction, as now must increase the size of the active subspace to reach the desired upper bound on the reduction errors. This work has provided a rigorous approach by which error bounds on the reduction errors can be established without having to execute the high fidelity model during the construction of the active subspace. The high fidelity model is only need to test the error bound using small number of executions. This methodology paves the way for a more efficient application of ROM to complex engineering models, where standard ROM techniques may be too taxing to optimally realize their value in practical situations. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗