Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Kernel”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2019

Evaluating LULESH Kernels on OpenCL FPGA

FPGAs are becoming promising heterogeneous computing components for high-performance computing. In this paper, we evaluate the resource utilizations, performance, and performance per watt of our implementations of the LULESH kernels in OpenCL on an Arria10-based FPGA platform. LULESH is a complex proxy application in the CORAL benchmark suite. We choose two representative kernels “CalcFBHourglassForceForElems” and “EvalEOSForElems” from the application in our study. Compared with the baseline implementations, our optimizations improve the performance by a factor of 1.65X and 2.96X for the two kernels on the FPGA, respectively. Using directives for accelerator programming, we also evaluate the performance of the kernels on an Intel Xeon 16-core CPU and an Nvidia K80 GPU. We find that the FPGA, constrained by the memory bandwidth, can perform 1.05X to 3.4X better than the CPU and GPU for small problem sizes. For the first kernel, the performance per watt on the FPGA is 1.59X and 7.1X higher than that on an Intel Xeon 16-core CPU and an Nvidia K80 GPU, respectively. For the second kernel, the performance per watt on the GPU is 1.82X higher than that on the FPGA. However, the performance per watt on the FPGA is 1.77X higher than that on the CPU.

FPGA↗
DOE OSTI2017

Kernel reconstruction methods for Doppler broadening — Temperature interpolation by linear combination of reference cross sections at optimally chosen temperatures

This paper establishes a new family of methods to perform temperature interpolation of nuclear interactions cross sections, reaction rates, or cross sections times the energy. One of these quantities at temperature T is approximated as a linear combination of quantities at reference temperatures (T j ). The problem is formalized in a cross section independent fashion by considering the kernels of the different operators that convert cross section related quantities from a temperature T 0 to a higher temperature T — namely the Doppler broadening operation. Doppler broadening interpolation of nuclear cross sections is thus here performed by reconstructing the kernel of the operation at a given temperature T by means of linear combination of kernels at reference temperatures (T j ). The choice of the L 2 metric yields optimal linear interpolation coefficients in the form of the solutions of a linear algebraic system inversion. The optimization of the choice of reference temperatures (T j ) is then undertaken so as to best reconstruct, in the L∞ sense, the kernels over a given temperature range [T min ,T max ]. The performance of these kernel reconstruction methods is then assessed in light of previous temperature interpolation methods by testing them upon isotope 238 U. Temperature-optimized free Doppler kernel reconstruction significantly outperforms all previous interpolation-based methods, achieving 0.1% relative error on temperature interpolation of 238 U total cross section over the temperature range [300 K,3000 K] with only 9 reference temperatures.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2019

Prediction of atomization energy using graph kernel and active learning

Data-driven prediction of molecular properties presents unique challenges to the design of machine learning methods concerning data structure/dimensionality, symmetry adaption, and confidence management. In this paper, we present a kernel-based pipeline that can learn and predict the atomization energy of molecules with high accuracy. The framework employs Gaussian process regression to perform predictions based on the similarity between molecules, which is computed using the marginalized graph kernel. To apply the marginalized graph kernel, a spatial adjacency rule is first employed to convert molecules into graphs whose vertices and edges are labeled by elements and interatomic distances, respectively. We then derive formulas for the efficient evaluation of the kernel. Specific functional components for the marginalized graph kernel are proposed, while the effects of the associated hyperparameters on accuracy and predictive confidence are examined. We show that the graph kernel is particularly suitable for predicting extensive properties because its convolutional structure coincides with that of the covariance formula between sums of random variables. Using an active learning procedure, we demonstrate that the proposed method can achieve a mean absolute error of 0.62 ± 0.01 kcal/mol using as few as 2000 training samples on the QM7 dataset.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2016

Technical Note: Image filtering to make computer-aided detection robust to image reconstruction kernel choice in lung cancer CT screening

Purpose: In lung cancer computed tomography (CT) screening, the performance of a computer-aided detection (CAD) system depends on the selection of the image reconstruction kernel. To reduce this dependence on reconstruction kernels, the authors propose a novel application of an image filtering method previously proposed by their group. Methods: The proposed filtering process uses the ratio of modulation transfer functions (MTFs) of two reconstruction kernels as a filtering function in the spatial-frequency domain. This method is referred to as MTF{sub ratio} filtering. Test image data were obtained from CT screening scans of 67 subjects who each had one nodule. Images were reconstructed using two kernels: f{sub STD} (for standard lung imaging) and f{sub SHARP} (for sharp edge-enhancement lung imaging). The MTF{sub ratio} filtering was implemented using the MTFs measured for those kernels and was applied to the reconstructed f{sub SHARP} images to obtain images that were similar to the f{sub STD} images. A mean filter and a median filter were applied (separately) for comparison. All reconstructed and filtered images were processed using their prototype CAD system. Results: The MTF{sub ratio} filtered images showed excellent agreement with the f{sub STD} images. The standard deviation for the difference between these images was very small, ∼6.0 Hounsfield units (HU). However, the mean and median filtered images showed larger differences of ∼48.1 and ∼57.9 HU from the f{sub STD} images, respectively. The free-response receiver operating characteristic (FROC) curve for the f{sub SHARP} images indicated poorer performance compared with the FROC curve for the f{sub STD} images. The FROC curve for the MTF{sub ratio} filtered images was equivalent to the curve for the f{sub STD} images. However, this similarity was not achieved by using the mean filter or median filter. Conclusions: The accuracy of MTF{sub ratio} image filtering was verified and the method was demonstrated to be effective for reducing the kernel dependence of CAD performance.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

SU-F-SPS-09: Parallel MC Kernel Calculations for VMAT Plan Improvement

Purpose: Adding kernels (small perturbations in leaf positions) to the existing apertures of VMAT control points may improve plan quality. We investigate the calculation of kernel doses using a parallelized Monte Carlo (MC) method. Methods: A clinical prostate VMAT DICOM plan was exported from Eclipse. An arbitrary control point and leaf were chosen, and a modified MLC file was created, corresponding to the leaf position offset by 0.5cm. The additional dose produced by this 0.5 cm × 0.5 cm kernel was calculated using the DOSXYZnrc component module of BEAMnrc. A range of particle history counts were run (varying from 3 × 10{sup 6} to 3 × 10{sup 7}); each job was split among 1, 10, or 100 parallel processes. A particle count of 3 × 10{sup 6} was established as the lower range because it provided the minimal accuracy level. Results: As expected, an increase in particle counts linearly increases run time. For the lowest particle count, the time varied from 30 hours for the single-processor run, to 0.30 hours for the 100-processor run. Conclusion: Parallel processing of MC calculations in the EGS framework significantly decreases time necessary for each kernel dose calculation. Particle counts lower than 1 × 10{sup 6} have too large of an error to output accurate dose for a Monte Carlo kernel calculation. Future work will investigate increasing the number of parallel processes and optimizing run times for multiple kernel calculations.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Atomistic and cluster dynamics modeling of fission gas (Xe) diffusivity in TRISO fuel kernels

TRISO fuel particles are candidates for use in next generation reactors including gas reactors, fluoride salt-cooled high temperature reactors, and micro-reactors. The UCO fuel kernel consists of a uranium dioxide (UO) and uranium carbide mixture. The addition of UC helps suppress the formation of carbon monoxide gas, which led to failures during initial TRISO development. The addition of uranium carbide alters the chemistry of the UO kernel, which is known to influence performance parameters such as fission gas diffusivity, although the impact has not been quantified and no models exist that take the change in chemistry into account. Therefore, better understanding and more accurate models of the impact of chemistry on fuel performance are of high priority. In this paper, a first-principles density functional theory (DFT) and empirical potential based multi-scale study has been carried out to model the diffusivity of fission gas xenon (Xe) in UCO TRISO fuel kernels. The focus is on the UO component in the UCO fuel kernels, as that represents the largest volume fraction of the fuel kernels. The study relies on DFT and empirical potential calculations to determine Xe and point defect properties, which are then used in thermodynamic and kinetic models to predict diffusion for intrinsic conditions. In addition, the information is utilized in cluster dynamics simulations using the Centipede code to estimate the impact of irradiation on defect transport. Additionally, the presence of UC or UC in the UCO fuel kernels is shown to have a substantial impact on the UO non-stoichiometry by inducing oxygen vacancies and driving UO sub-stoichiometric, which causes much slower Xe diffusion in UCO compared to light water reactor UO fuel. The application of this model in fuel performance simulations using the Bison code is also demonstrated.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2018

Phase quantification in low carbon Nb-Mo bearing steel by electron backscatter diffraction technique coupled with kernel average misorientation

Highlights: • The microstructures in a low carbon Nb-Mo bearing steel have been investigated. • EBSD combined with a (KAM) map has been used for quantification. • The effects of step size and kernel size on quantification are systemically discussed. • The result indicates that the kernel size should be close to the sub-structure of the granular bainite. - Abstract: An efficient and accurate method is developed for microstructural quantification of complex phases in a low carbon Nb-Mo bearing steel, of which optical micrographs show that it consists of granular bainite and a small amount of ferrite. Our previous work, proposing a method to measure the misorientation angles via electron backscatter diffraction (EBSD) to differentiate granular bainite and ferrite, has been reported. That method is accurate for phase quantification but laborious for the characterization process. To resolve this difficulty, in this study, EBSD combined with a kernel average misorientation (KAM) map is used for phase characterization. Comparisons are made among KAM maps with different kernel sizes (300 to 600 nm) and various step sizes (100 to 600 nm). It is found that a kernel size close to the sub-structure size of granular bainite (500 nm) is optimal for phase identification, while varied step sizes produce relatively invariant results. Therefore, KAM maps can be used for fast and reliable phase quantification, provided that an appropriate kernel size and a large step size are used.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Progress towards understanding ultranonlocality through the wave-vector and frequency dependence of approximate exchange-correlation kernels

In the framework of time-dependent density functional theory (TDDFT), the exact exchange-correlation (xc) kernel f xc (n, q, ω) determines the ground-state energy, excited-state energies, lifetimes, and the time-dependent linear density response of any many-electron system. The recently developed MCP07 xc kernel f xc (n, q, ω) of A. Ruzsinszky et al. [Phys. Rev. B 101, 245135 (2020)] yields excellent uniform electron gas (UEG) ground-state energies and plausible plasmon lifetimes. As MCP07 is constructed to describe f xc of the UEG, it cannot capture optical properties of real materials. To verify this claim, we follow Nazarov et al. [Phys. Rev. Lett. 102, 113001 (2009)] to construct the long-range, dynamic xc kernel, lim q→0 f xc (n, q, ω) = -α(ω)e 2 /q 2 , of a weakly inhomogeneous electron gas, using MCP07 and other common xc kernels. The strong wavevector and frequency dependence of the “ultranonlocality” coefficient α(ω) is demonstrated for a variety of simple metals and semiconductors. We examine how imposing exact constraints on an approximate kernel shapes α(ω). Comparisons to kernels derived from correlated-wavefunction calculations are drawn.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Advanced stationary and nonstationary kernel designs for domain-aware Gaussian processes

Gaussian process regression is a widely-applied method for function approximation and uncertainty quantification. The technique has gained popularity recently in the machine learning community due to its robustness and interpretability. The mathematical methods we discuss in this paper are an extension of the Gaussian-process framework. We are proposing advanced kernel designs that only allow for functions with certain desirable characteristics to be elements of the reproducing kernel Hilbert space (RKHS) that underlies all kernel methods and serves as the sample space for Gaussian process regression. These desirable characteristics reflect the underlying physics; two obvious examples are symmetry and periodicity constraints. In addition, non-stationary kernel designs can be defined in the same framework to yield flexible multi-task Gaussian processes. We will show the impact of advanced kernel designs on Gaussian processes using several synthetic and two scientific data sets. The results of our research show that including domain knowledge, communicated through advanced kernel designs, has a significant impact on the accuracy and relevance of the function approximation.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2019

A Case Study with the HACCmk Kernel in SYCL

The SYCL standard specifies a cross-platform abstraction layer that enables programming of heterogeneous computing systems using standard C++ [1]. In the Open Computing Language (OpenCL) programming model, host and device codes are written in different languages [2 ]. To improve development productivity and performance portability, the SYCL programming model combines host and device codes for an application in a type-safe way. In this report, we are interested in applying the SYCL programming model to a computationally intensive routine derived from the Hardware Accelerated Cosmology Code (HACC) framework for a study on performance portability on a heterogeneous computing device. We use Intel® OneAPI toolkit [3] to build the OpenCL and SYCL programs, and evaluate the performance of both implementations on Intel® integrated GPUs. We find that the SYCL implementation can achieve the same performance as the OpenCL implementation after we specify a target backend for the compiler to build the SYCL program in offline compilation. Without the specification of a target backend, the runtime will have to compile intermediate device codes for a target platform. Such runtime overhead may become significant compared to the execution of a kernel on a target device. As the kernel routine is compute-bound, we evaluate the impact of the number of compute units upon the kernel’s raw performance using two GPUs. When the hardware resource is not underutilized, we can obtain almost linear performance speedup from 48 compute units to 72 compute units in offline compilation, which indicates that the number of compute units on a GPU are important to improving the raw performance of a compute-bound kernel. The experimental results show that SYCL is a promising programming model for heterogeneous computing. The remainder of the report is organized as follows: In Section II, we describe the SYCL programming model, the steps to map an OpenCL program to a SYCL program, and the architecture of an integrated GPU. Section III introduces the kernel routine, and presents the OpenCL and SYCL implementations of the kernel. In Section IV, we evaluate the performance of the implementations on the GPUs. Section V concludes the report.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2017

ECP Milestone Report: Identify initial kernels, bake-off problems (benchmarks) and miniapps (CEED-MS6)

The CEED co-design center is developing a number of kernels, bake-off / benchmark problems (BPs) and mini-applications (miniapps) that capture the unique requirements of high-order finite element algorithms and use them to influence standards, best practices and vendors to generate more portable, better performing code across diverse hardware designs. This is a main activity in CEED’s Hardware, Applications and Finite Element thrusts. CEED’s miniapps, a.k.a. CEEDlings, are small, standalone applications - surrogates for sub-components of interest from the selected first wave of applications (the MARBL/LLNLApp and ExaSMR projects). The BPs are even smaller standalone programs that solve a simple physical problem, including discretization, linear and/or nonlinear solve, and potentially a relatively small number of time steps. A BP may be a proxy/surrogate for part of a miniapp or another problem of general interest. In CEED, both miniapps and BPs will be used extensively in the project to evaluate and compare the performance of algorithms in specific contexts. The parts of the miniapps and BPs that play an important role in their overall performance will be identified and separated as standalone kernels. These kernels and their interaction inside BPs, miniapps and the actual applications, will be the main focus of our optimization efforts. All three standalone pieces of software (miniapps, BPs, and kernels) will serve as a basis for engaging and collaborating with hardware vendors and software technologies (ST) projects. These partners can start at the bottom of the hierarchy, e.g. with the kernels, and easily transition to BPs and miniapps as all three families will be produced in the same CEED environment. In this milestone we identified the initial kernels, bake-off problems and miniapps and delivered software and documentation for them through the new CEED website, http://ceed.exascaleproject.org. We specifically identified the Nekbone miniapp for the ExaSMR application and developed a new miniapp, Laghos, for the MARBL/LLNLApp application. We also formulated four initial benchmark problems (BP1- BP4) with clearly defined performance metrics and performed an initial bake-off comparison between the Nek5000 (spectral element) and MFEM (high-order finite element) technologies. As part of the milestone, we also engaged vendors (AMD) and ST projects (MPICH, STRUMPACK). In this document we are reporting some details and results from these R&D activities, as well as additional project-wide activities performed in Q3 of FY17, including: making CEED software available on GitHub, developing the project website, and the preparation of CEED’s first annual meeting.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Clutter identification based on kernel density estimation and sparse recovery

A cognitive radar framework is being developed to dynamically detect changes in the clutter characteristics, and to adapt to these changes by identifying the new clutter distribution. In our previous work, we have presented a sparse-recovery based clutter identification technique. In this technique, each column of the dictionary represents a specific distribution. More specifically, calibration radar clutter data corresponding to a specific distribution is transformed into a distribution through kernel density estimation. When the new batch of radar data arrives, the new data is transformed to a distribution through the same kernel density estimation method and its distribution characteristics is identified through sparse-recovery. In this paper, we extend our previous work to consider different kernels and kernel parameters for sparse-recovery-based clutter identification and the numerical results are presented as well. The impact of different kernels and kernel parameters are analyzed by comparing the identification accuracy of each scenario.

Wang, Haokun↗
DOE OSTI2024

Parallel hybrid quantum-classical machine learning for kernelized time-series classification

Supervised time-series classification garners widespread interest because of its applicability throughout a broad application domain including finance, astronomy, biosensors, and many others. Here, in this work, we tackle this problem with hybrid quantum-classical machine learning, deducing pairwise temporal relationships between time-series instances using a timeseries Hamiltonian kernel (TSHK). A TSHK is constructed with a sum of inner products generated by quantum states evolved using a parameterized time evolution operator. This sum is then optimally weighted using techniques derived from multiple kernel learning. Because we treat the kernel weighting step as a differentiable convex optimization problem, our method can be regarded as an end-to-end learnable hybrid quantum-classical-convex neural network, or QCC-net, whose output is a data set-generalized kernel function suitable for use in any kernelized machine learning technique such as the support vector machine (SVM). Using our TSHK as input to a SVM, we classify univariate and multivariate time-series using quantum circuit simulators and demonstrate the efficient parallel deployment of the algorithm to 127-qubit superconducting quantum processors using quantum multi-programming.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Simultaneous dissection of grain carotenoid levels and kernel color in biparental maize populations with yellow-to-orange grain

Maize enriched in provitamin A carotenoids could be key in combatting vitamin A deficiency in human populations relying on maize as a food staple. Consumer studies indicate that orange maize may be regarded as novel and preferred. This study identifies genes of relevance for grain carotenoid concentrations and kernel color, through simultaneous dissection of these traits in 10 families of the US maize nested association mapping panel that have yellow to orange grain. Quantitative trait loci were identified via joint-linkage analysis, with phenotypic variation explained for individual kernel color quantitative trait loci ranging from 2.4% to 17.5%. These quantitative trait loci were cross-analyzed with significant marker-trait associations in a genome-wide association study that utilized ~27 million variants. Nine genes were identified: four encoding activities upstream of the core carotenoid pathway, one at the pathway branchpoint, three within the α- or β-pathway branches, and one encoding a carotenoid cleavage dioxygenase. Of these, three exhibited significant pleiotropy between kernel color and one or more carotenoid traits. Kernel color exhibited moderate positive correlations with β-branch and total carotenoids and negligible correlations with α-branch carotenoids. These findings can be leveraged to simultaneously achieve desirable kernel color phenotypes and increase concentrations of provitamin A and other priority carotenoids.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Compactly‐Supported Nonstationary Kernels for Computing Exact Gaussian Processes on Big Data

The Gaussian process (GP) is a widely used method for analyzing large-scale data sets, including spatio-temporal measurements of nonlinear processes that are now commonplace in the environmental sciences. Traditional implementations of GPs involve stationary kernels (also termed covariance functions) that limit their flexibility, and exact methods for inference that prevent application to data sets with more than about 10,000 points. Modern approaches to address stationarity assumptions generally fail to accommodate large data sets, while all attempts to address scalability focus on approximating the Gaussian likelihood, which can involve subjectivity and lead to inaccuracies. In this work, we explicitly derive an alternative kernel that can discover and encode both sparsity and nonstationarity. We embed the kernel within a fully Bayesian GP model and leverage high-performance computing resources to enable the analysis of massive data sets. We demonstrate the favorable performance of our novel kernel relative to existing exact and approximate GP methods across a variety of synthetic data examples. Furthermore, we conduct space–time prediction based on more than 1 million measurements of daily maximum temperature and verify that our results outperform state-of-the-art methods in the Earth sciences. More broadly, having access to exact GPs that use ultra-scalable, sparsity-discovering, nonstationary kernels allows GP methods to truly compete with a wide variety of machine learning methods.

Gaussian processes↗
DOE OSTI2019

Using Compiler Directives for Performance Portability in Scientific Computing: Kernels from Molecular Simulation

Achieving performance portability for high-performance computing (HPC) applications in scientific fields has become an increasingly important initiative due to large differences in emerging supercomputer architectures. Here we test some key kernels from molecular dynamics (MD) to determine whether the use of the OpenACC directive-based programming model when applied to these kernels can result in performance within an acceptable range for these types of programs in the HPC setting. We find that for easily parallelizable kernels, performance on the GPU remains within this range. On the CPU, OpenACC-parallelized pairwise distance kernels would not meet the performance standards required, when using AMD Opteron “Interlagos” processors, but with IBM Power 9 processors, performance remains within an acceptable range for small batch sizes. These kernels provide a test for achieving performance portability with compiler directives for problems with memory-intensive components as are often found in scientific applications.

Sedova, Ada↗
DOE OSTI2026

Inverse Mapping of the Collision Kernel and Wall Flux Scaling in a Tall Convection‐Cloud Chamber Using Local Sensors and Knowledge‐Informed Deep Learning

Droplet collision–coalescence is a crucial process in cloud physics, but accurately representing this process under different dynamical conditions remains challenging. A proposed future convective‐cloud chamber aims to investigate this key process, but the method for observing it remains unclear, even though it is theoretically established that collision‐coalescence will occur. This study serves as a proof‐of‐concept demonstration of how knowledge‐informed deep learning, combined with measurement data from local sensors in the chamber, can be used to estimate the collision kernels, which determine how the droplet size distribution evolves during collision‐coalescence. In addition to estimating the collision kernel, we also address wall fluxes, another uncertain but important process that acts as a source of heat and moisture in the chamber. Ensemble runs of large‐eddy simulations are conducted by scaling the wall fluxes and the collision kernel, while the measured flow and cloud properties are used as inputs for a neural network. Results indicate that this approach successfully maps the scaling of wall fluxes and the collision kernel with biases of approximately 1% or less relative to the range of the target data. This proof‐of‐concept lays the groundwork for future applications; when the real measurements are available, real sensor data combined with the trained model presented in this work will enable estimation of the actual wall fluxes and collision kernel.

cloud chamber↗
DOE OSTI2023

Exact Gaussian processes for massive datasets via non-stationary sparsity-discovering kernels

Abstract A Gaussian Process (GP) is a prominent mathematical framework for stochastic function approximation in science and engineering applications. Its success is largely attributed to the GP’s analytical tractability, robustness, and natural inclusion of uncertainty quantification. Unfortunately, the use of exact GPs is prohibitively expensive for large datasets due to their unfavorable numerical complexity of $$O(N^3)$$ O ( N 3 ) in computation and $$O(N^2)$$ O ( N 2 ) in storage. All existing methods addressing this issue utilize some form of approximation—usually considering subsets of the full dataset or finding representative pseudo-points that render the covariance matrix well-structured and sparse. These approximate methods can lead to inaccuracies in function approximations and often limit the user’s flexibility in designing expressive kernels. Instead of inducing sparsity via data-point geometry and structure, we propose to take advantage of naturally-occurring sparsity by allowing the kernel to discover—instead of induce—sparse structure. The premise of this paper is that the data sets and physical processes modeled by GPs often exhibit natural or implicit sparsities, but commonly-used kernels do not allow us to exploit such sparsity. The core concept of exact, and at the same time sparse GPs relies on kernel definitions that provide enough flexibility to learn and encode not only non-zero but also zero covariances. This principle of ultra-flexible, compactly-supported, and non-stationary kernels, combined with HPC and constrained optimization, lets us scale exact GPs well beyond 5 million data points.

97 MATHEMATICS AND COMPUTING↗