Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Arithmetic”

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 235 records · Page 13

NgramPPM: Compression Analytics without Compression

Arithmetic Coding (AC) using Prediction by Partial Matching (PPM) is a compression algorithm that can be used as a machine learning algorithm. This paper describes a new algorithm, NGram PPM. NGram PPM has all the predictive power of AC/PPM, but at a fraction of the computational cost. Unlike compression-based analytics, it is also amenable to a vector space interpretation, which creates the ability for integration with other traditional machine learning algorithms. AC/PPM is reviewed, including its application to machine learning. Then NGram PPM is described and test results are presented, comparing them to AC/PPM.

97 MATHEMATICS AND COMPUTING↗

Homomorphic Encryption for Machine Learning and Artificial Intelligence Applications

Third-party and expert analysis is a cost-effective solution for solving specialized problems or processing large datasets related to reactor structural health monitoring and nondestructive evaluation. However, when handling proprietary information, third-party and expert analysts pose a privacy risk. To address this challenge, Homomorphic Encryption (HE) permits arithmetic operations on encrypted data without exposing the underlying data. Implementations of Machine Learning (ML) and Artificial Intelligence (AI) algorithms using HE greatly enhances the capabilities of third-party analysts while maintaining a low security risk. This paper details current success in applying Principal Component Analysis (PCA) and Fully Connected Neural Networks (NN) using the Microsoft SEAL implementation of the popular CKKS Fully Homomorphic Encryption (FHE) algorithm. The MNIST Handwritten Dataset is analyzed as a proof-of-concept demonstration of the implementations.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Reconfigurable neuromorphic components and algorithms for next-generation artificial intelligence

Digital transistor-based general-purpose hardware (e.g., central processing units) is the dominant solution to support both traditional computing (logic, arithmetic, etc.) as well as modern artificial intelligence. State-of-the-art research has shown feasibility of post-digital physics-based neuromorphic hardware, which is hypothesized to support artificial intelligence algorithms with orders-of-magnitude improved time/energy efficiencies. But such research has not been widely deployed mainly because of such novel hardware’s extreme application-specificity, and the dominance of low-cost general-purpose (but inefficient) digital hardware. To make use of the novel algorithms and the superlative performance of physics-based hardware, we need to identify scientific principles that can enable generality in physics-based hardware. This work resulted in two important broad outcomes – first, we demonstrate fully reconfigurable neuromorphic components, and second, we demonstrate a viable artificial intelligence learning algorithm that can exploit the functioning of neuromorphic hardware. We demonstrate up to five orders of magnitude improvement in energy efficiency compared to the best general-purpose digital hardware.

97 MATHEMATICS AND COMPUTING↗

AI and Machine Learning with Encrypted Data (CRADA Final Report)

The proposed project aimed to develop faster algorithms for FHE and zk-SNARKs processing. The research team previously developed a fast Number Theoretic Transform (NTT) and other fast arithmetic algorithms. The project team explored using these fast algorithms to enable FHE and zk-SNARKSs acceleration, with specific attention to how the algorithms would need to be integrated into specific FHE and zk-SNARKs software packages to become useful. Existing FHE/zk-SNARKs software packages have been written for CPUs, however this project’s technology targeted FPGAs and GPUs. The team explored how to modify existing CPU software to support FPGAs/GPUs as well as ways they could implement entirely new software stacks that run purely on FPGAs/GPUs.

97 MATHEMATICS AND COMPUTING↗

Experiences with SYCL on AMD GPUs with Kokkos

With the recent diversification of the hardware landscape in the high-performance computing (HPC) community, performance-portability solutions are becoming more and more important. One of the most popular choices is Kokkos, which recently became a Linux Foundation project. Most of its development is supported by the US Department of Energy and the French Alternative Energies and Atomic Energy Commission. Kokkos is implemented as a C++ library with multiple backends to support CPUs as well as various GPU architectures. These backends include OpenMP, CUDA, HIP, and also SCYL. This approach enables users to leverage the preferred vendor toolchain for the respective platform (e.g. CUDA, ROCm, OneAPI). The SYCL backend is used to target Intel GPUs, in particular to support the Aurora exascale supercomputer. However, SYCL itself also offers a large degree of portability, and in fact Kokkos’ CI for SYCL has been running on NVIDIA hardware due to a lack of access to Intel GPUs. In this report, we describe our experience with using Kokkos SYCL backend on AMD GPUs targeting the Frontier supercomputer at Oak Ridge National Laboratory. The two major SYCL implementations are DPC++ and AdaptiveCpp. While the Kokkos SYCL backend has been implemented using the former, the latter was the first implementation to target AMD GPUs. We will discuss the experience with both of these SYCL implementations in terms of functionality and performance. Using Kokkos to evaluate SYCL toolchains has a number of benefits. Kokkos’ use of SYCL is fairly complex, exercising features such as graphs, relocatable device functions, atomics – including for non-arithmetic types, as well as pinned and page migratable memory allocations. Kokkos also needs to implement capabilities such as Kokkos’ hierarchical parallelism that are not a straight-forward mapping to SYCL capabilities. Furthermore, a large number of libraries and applications that represent diverse use cases are implemented in Kokkos, providing readily available test cases for a toolchain evaluation. Preliminary results show that support for AMD GPUs in DPC++ is much less mature than for NVIDIA GPUs or Intel GPUs. While the situation has improved significantly over the last year, we still encounter many runtime failures, dispatching problems, and code generation issues. With AdaptiveCpp the challenges arise even earlier in the evaluation process. Since Kokkos’ SYCL implementation is largely focused on supporting Intel GPUs, we opted to leverage SYCL extensions which are available in DPC++ but not in AdaptiveCpp. Furthermore, AdaptiveCpp appears to be less conformant with the SYCL2020 standard which Kokkos relies on. In some cases, we are able to work around the lack of feature support, in other cases we have to disable certain Kokkos capabilities to evaluate the toolchain. Our evaluation will leverage Kokkos’ unit tests to establish basic functionality and feature completeness. We then use simple benchmarks for components of a CG implementation as a measure of usability and performance of the SYCL toolchains.

97 MATHEMATICS AND COMPUTING↗

Controlling Oxidation of Nb in Oxygen Abundant Environments

Modern particle accelerators depend on Superconducting Radio Frequency (SRF) cavities made from high-purity niobium (Nb) to achieve optimal performance, including high quality factors and strong accelerating gradients. However, when exposed to air, niobium naturally forms a complex oxide layer that can introduce surface imperfections and carbon contamination. This work examines an alternative oxidation strategy under oxygen-rich conditions to better regulate the oxide formation process. The ultimate objective is to improve surface uniformity and cleanliness, thereby reducing defect density and enhancing performance. We used Confocal Microscopy, Scanning Electron Microscopy (SEM), and X-Ray Photoelectron Spectroscopy (XPS) to analyze surface changes. We used standard metrics, like Arithmetic Average Roughness (Ra) and Root Mean Square (Rq). Three oxidation methods were applied: short, extended, and a repeated HF and H₂O₂ oxidation process. Preliminary results show promising reductions in carbon contamination and surface defects.

Romero, Juan [Fermilab]↗

The Greatest Mathematical Discovery?

What mathematical discovery more than 1500 years ago: (1) Is one of the greatest, if not the greatest, single discovery in the field of mathematics? (2) Involved three subtle ideas that eluded the greatest minds of antiquity, even geniuses such as Archimedes? (3) Was fiercely resisted in Europe for hundreds of years after its discovery? (4) Even today, in historical treatments of mathematics, is often dismissed with scant mention, or else is ascribed to the wrong source? Answer: Our modern system of positional decimal notation with zero, together with the basic arithmetic computational schemes, which were discovered in India about 500 CE.

97 MATHEMATICS AND COMPUTING↗

Infinite quantum signal processing

Quantum signal processing (QSP) represents a real scalar polynomial of degree d using a product of unitary matrices of size 2 × 2 , parameterized by ( d + 1 ) real numbers called the phase factors. This innovative representation of polynomials has a wide range of applications in quantum computation. When the polynomial of interest is obtained by truncating an infinite polynomial series, a natural question is whether the phase factors have a well defined limit as the degree d → ∞ . While the phase factors are generally not unique, we find that there exists a consistent choice of parameterization so that the limit is well defined in the ℓ 1 space. This generalization of QSP, called the infinite quantum signal processing, can be used to represent a large class of non-polynomial functions. Our analysis reveals a surprising connection between the regularity of the target function and the decay properties of the phase factors. Our analysis also inspires a very simple and efficient algorithm to approximately compute the phase factors in the ℓ 1 space. The algorithm uses only double precision arithmetic operations, and provably converges when the ℓ 1 norm of the Chebyshev coefficients of the target function is upper bounded by a constant that is independent of d . This is also the first numerically stable algorithm for finding phase factors with provable performance guarantees in the limit d → ∞ .

Dong, Yulong [Department of Mathematics, Universit↗

Similarity of the turbulent kinetic energy dissipation rate distribution in the upper mixed layer of the tropical Indian Ocean

Turbulence within the upper ocean mixed layer plays a key role in various physical, biological, and chemical processes. Between September and November 2011, a dataset of 570 vertical profiles of the turbulent kinetic energy (TKE) dissipation rate, as well as conventional hydrological and meteorological data, were collected in the upper layer of the tropical Indian Ocean. These data were used to statistically analyze the vertical distribution of the TKE dissipation rate in the mixed layer. The arithmetic-mean method made the statistical TKE dissipation rate profile more scattered than the median and geometric-mean methods. The statistical TKE dissipation rate were respectively scaled by the surface buoyancy flux and the TKE dissipation rate at the mixed-layer base. It was found that the TKE dissipation rate scaled by that at the mixed-layer base exhibited better similarity characteristics than that scaled by the surface buoyancy flux, whether the stability parameter D/|L MO | was greater or less than 10, indicating that the TKE dissipation rate at the mixed-layer base is a better characteristic scaling parameter for reflecting the intrinsic structure of the TKE dissipation rate in the mixed layer, where D and L MO are respectively the mixed-layer thickness and the Monin-Obukhov length scale. The parameterization of the TKE dissipation rate at the mixed-layer base on the shear-driven dissipation rate and the surface buoyancy flux was further explored. It was found that the TKE dissipation rate at the mixed-layer base could be well fitted by a linear combination of three terms: the wind-shear-driven dissipation rate, the surface buoyancy flux, and a simple nonlinear coupling term of these two.

54 ENVIRONMENTAL SCIENCES↗

EMUlator: An Elementary Metabolite Unit (EMU) Based Isotope Simulator Enabled by Adjacency Matrix

Stable isotope based metabolic flux analysis is currently the unique methodology that allows the experimental study of the integrated responses of metabolic networks. This method primarily relies on isotope labeling and modeling, which could be a challenge in both experimental and computational biology. Specifically, the algorithm implementation for isotope simulation is a critical step, limiting extensive usage of this powerful approach. In this work, we introduce EMUlator a Python-based isotope simulator which is developed on Elementary Metabolite Unit (EMU) algorithm, an efficient and powerful algorithm for isotope modeling. We propose a novel adjacency matrix method to implement EMU modeling and exemplify it stepwise. This method is intuitively straightforward and can be conveniently mastered for various customized purposes. We apply this arithmetic pipeline to understand the phosphoketolase flux in the metabolic network of an industrial microbe Clostridium acetobutylicum . The resulting design enables a high-throughput and non-invasive approach for estimating phosphoketolase flux in vivo . Our computational insights allow the systematic design and prediction of isotope-based metabolic models and yield a comprehensive understanding of their limitations and potentials.

09 BIOMASS FUELS↗

Making Steppingstones out of Stumbling Blocks: A Bayesian Model Evidence Estimator with Application to Groundwater Transport Model Selection

Bayesian model evidence (BME) is a measure of the average fit of a model to observation data given all the parameter values that the model can assume. By accounting for the trade-off between goodness-of-fit and model complexity, BME is used for model selection and model averaging purposes. For strict Bayesian computation, the theoretically unbiased Monte Carlo based numerical estimators are preferred over semi-analytical solutions. This study examines five BME numerical estimators and asks how accurate estimation of the BME is important for penalizing model complexity. The limiting cases for numerical BME estimators are the prior sampling arithmetic mean estimator (AM) and the posterior sampling harmonic mean (HM) estimator, which are straightforward to implement, yet they result in underestimation and overestimation, respectively. We also consider the path sampling methods of thermodynamic integration (TI) and steppingstone sampling (SS) that sample multiple intermediate distributions that link the prior and the posterior. Although TI and SS are theoretically unbiased estimators, they could have a bias in practice arising from numerical implementation. For example, sampling errors of some intermediate distributions can introduce bias. We propose a variant of SS, namely the multiple one-steppingstone sampling (MOSS) that is less sensitive to sampling errors. We evaluate these five estimators using a groundwater transport model selection problem. SS and MOSS give the least biased BME estimation at an efficient computational cost. If the estimated BME has a bias that covariates with the true BME, this would not be a problem because we are interested in BME ratios and not their absolute values. On the contrary, the results show that BME estimation bias can be a function of model complexity. Thus, biased BME estimation results in inaccurate penalization of more complex models, which changes the model ranking. This was less observed with SS and MOSS as with the three other methods.

54 ENVIRONMENTAL SCIENCES↗

Quantum modularity and complex Chern-Simons theory

The Quantum Modularity Conjecture of Zagier predicts the existence of a formal power series with arithmetically interesting coefficients that appears in the asymptotics of the Kashaev invariant at each root of unity. Our goal is to construct a power series from a Neumann-Zagier datum (i.e., an ideal triangulation of the knot complement and a geometric solution to the gluing equations) and a complex root of unity $\zeta$. We prove that the coefficients of our series lie in the trace field of the knot, adjoined a complex root of unity. We conjecture that our series are those that appear in the Quantum Modularity Conjecture and confirm that they match the numerical asymptotics of the Kashaev invariant (at various roots of unity) computed by Zagier and the first author. Our construction is motivated by the analysis of singular limits in Chern-Simons theory with gauge group $SL(2,C)$ at fixed level $k$, where $\zeta^k=1$.

97 MATHEMATICS AND COMPUTING↗

HGQ: High Granularity Quantization for Real-time Neural Networks on FPGAs

Neural networks with sub-microsecond inference latency are required by many critical applications. Targeting such applications deployed on FPGAs, we present High Granularity Quantization (HGQ), a quantization-aware training framework that optimizes parameter bit-widths through gradient descent. Unlike conventional methods, HGQ determines the optimal bit-width for each parameter independently, making it suitable for hardware platforms supporting heterogeneous arbitrary precision arithmetic. In our experiments, HGQ shows superior performance compared to existing network compression methods, achieving orders of magnitude reduction in resource consumption and latency while maintaining the accuracy on several benchmark tasks. These improvements enable the deployment of complex models previously infeasible due to resource or latency constraints. HGQ is open-source and is used for developing next-generation trigger systems at the CERN ATLAS and CMS experiments for particle physics, enabling the use of advanced machine learning models for real-time data selection with sub-microsecond latency.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Polynomial time blackbox identity testers for depth-3 circuits : the field doesn't matter.

Let C be a depth-3 circuit with n variables, degree d and top fanin k (called {Sigma}{Pi}{Sigma}(k, d, n) circuits) over base field F. It is a major open problem to design a deterministic polynomial time blackbox algorithm that tests if C is identically zero. Klivans & Spielman (STOC 2001) observed that the problem is open even when k is a constant. This case has been subjected to a serious study over the past few years, starting from the work of Dvir & Shpilka (STOC 2005). We give the first polynomial time blackbox algorithm for this problem. Our algorithm runs in time poly(n)d{sup k}, regardless of the base field. The only field for which polynomial time algorithms were previously known is F = Q (Kayal & Saraf, FOCS 2009, and Saxena & Seshadhri, FOCS 2010). This is the first blackbox algorithm for depth-3 circuits that does not use the rank based approaches of Karnin & Shpilka (CCC 2008). We prove an important tool for the study of depth-3 identities. We design a blackbox polynomial time transformation that reduces the number of variables in a {Sigma}{Pi}{Sigma}(k, d, n) circuit to k variables, but preserves the identity structure. Polynomial identity testing (PIT) is a major open problem in theoretical computer science. The input is an arithmetic circuit that computes a polynomial p(x{sub 1}, x{sub 2},..., x{sub n}) over a base field F. We wish to check if p is the zero polynomial, or in other words, is identically zero. We may be provided with an explicit circuit, or may only have blackbox access. In the latter case, we can only evaluate the polynomial p at various domain points. The main goal is to devise a deterministic blackbox polynomial time algorithm for PIT.

99 GENERAL AND MISCELLANEOUS↗

A communication-avoiding, hybrid-parallel, rank-revealing orthogonalization method.

Orthogonalization consumes much of the run time of many iterative methods for solving sparse linear systems and eigenvalue problems. Commonly used algorithms, such as variants of Gram-Schmidt or Householder QR, have performance dominated by communication. Here, 'communication' includes both data movement between the CPU and memory, and messages between processors in parallel. Our Tall Skinny QR (TSQR) family of algorithms requires asymptotically fewer messages between processors and data movement between CPU and memory than typical orthogonalization methods, yet achieves the same accuracy as Householder QR factorization. Furthermore, in block orthogonalizations, TSQR is faster and more accurate than existing approaches for orthogonalizing the vectors within each block ('normalization'). TSQR's rank-revealing capability also makes it useful for detecting deflation in block iterative methods, for which existing approaches sacrifice performance, accuracy, or both. We have implemented a version of TSQR that exploits both distributed-memory and shared-memory parallelism, and supports real and complex arithmetic. Our implementation is optimized for the case of orthogonalizing a small number (5-20) of very long vectors. The shared-memory parallel component uses Intel's Threading Building Blocks, though its modular design supports other shared-memory programming models as well, including computation on the GPU. Our implementation achieves speedups of 2 times or more over competing orthogonalizations. It is available now in the development branch of the Trilinos software package, and will be included in the 10.8 release.

99 GENERAL AND MISCELLANEOUS↗

Study of Bbar to Xu l nubar Decays in BBbar Events Tagged by a Fully Reconstructed B-meson Decay and Determination of |V_{ub}|

We report measurements of partial branching fractions for inclusive charmless semileptonic B decays {bar B} {yields} X{sub u}{ell}{bar {nu}}, and the determination of the CKM matrix element |V{sub ub}|. The analysis is based on a sample of 467 million {Upsilon}(4S) {yields} B{bar B} decays recorded with the BABAR detector at the PEP-II e{sup +}e{sup -} storage rings. We select events in which the decay of one of the B mesons is fully reconstructed and an electron or a muon signals the semileptonic decay of the other B meson. We measure partial branching fractions {Delta}{Beta} in several restricted regions of phase space and determine the CKM element |V{sub ub}| based on four different QCD predictions. For decays with a charged lepton momentum p*{sub {ell}} > 1.0 GeV in the B meson rest frame, we obtain {Delta}{Beta} = (1.80 {+-} 0.13{sub stat.} {+-} 0.15{sub sys.} {+-} 0.02{sub theo.}) x 10{sup -3} from a maximum likelihood fit to the two-dimensional M{sub X} - q{sup 2} distribution. Here, M{sub X} refers to the invariant mass of the final state hadron X and q{sup 2} is the invariant mass squared of the charged lepton and neutrino. From this measurement we extract |V{sub ub}| = (4.31 {+-} 0.25{sub exp.} {+-} 0.16{sub theo.}) x 10{sup -3} as the arithmetic average of four results obtained from four different QCD predictions of the partial rate. We separately determine partial branching fractions for {bar B}{sup 0} and B{sup -} decays and derive a limit on the isospin breaking in {bar B} {yields} X{sub u}{ell}{bar {nu}} decays.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

High-order entropy-based closures for linear transport in slab geometry II: A computational study of the optimization problem

We present a numerical algorithm to implement entropy-based (M{sub N}) moment models in the context of a simple, linear kinetic equation for particles moving through a material slab. The closure for these models - as is the case for all entropy-based models - is derived through the solution of constrained, convex optimization problem. The algorithm has two components. The first component is a discretization of the moment equations which preserves the set of realizable moments, thereby ensuring that the optimization problem has a solution (in exact arithmetic). The discretization is a second-order kinetic scheme which uses MUSCL-type limiting in space and a strong-stability-preserving, Runge-Kutta time integrator. The second component of the algorithm is a Newton-based solver for the dual optimization problem, which uses an adaptive quadrature to evaluate integrals in the dual objective and its derivatives. The accuracy of the numerical solution to the dual problem plays a key role in the time step restriction for the kinetic scheme. We study in detail the difficulties in the dual problem that arise near the boundary of realizable moments, where quadrature formulas are less reliable and the Hessian of the dual objection function is highly ill-conditioned. Extensive numerical experiments are performed to illustrate these difficulties. In cases where the dual problem becomes 'too difficult' to solve numerically, we propose a regularization technique to artificially move moments away from the realizable boundary in a way that still preserves local particle concentrations. We present results of numerical simulations for two challenging test problems in order to quantify the characteristics of the optimization solver and to investigate when and how frequently the regularization is needed.

99 GENERAL AND MISCELLANEOUS↗

Reliable algorithms for power system analysis in the presence of data uncertainties

A robust and reliable power flow analysis represents an essential requirement for many power systems applications as far as network optimization, voltage control, state estimation, and service restoration are concerned. The most common power flow approach, referred to here as a deterministic power flow (PLF), requires precise or 'crisp' values chosen by the analyst for each input variable. The solution provides precise network voltages and flows through each line. The specified values rest upon assumptions about the operating condition derived from historical measurements or predictions about future conditions and thus, cannot be considered accurate. Even in the case where the inputs are based on measurements, inaccuracies arise from time-skew problems, three-phase unbalance, static modeling approximations of dynamic components (e.g., transformer tap changers), variations in line parameters, and so on. The advent of deregulation and competitive power markets will only exacerbate this problem as well-known generation patterns change, loading becomes less predictable and the transmission paths grow more diverse. Conventional methodologies proposed in literature address tolerance analysis of power flow solution by means of detailed probabilistic methods, accounting for the variability and stochastic nature of the input data, and sampling based approaches. In particular uncertainty propagation using sampling based methods, such as the Monte Carlo, requires several model runs that sample various combinations of input values. Since the number of model runs can sometimes be very large, the required computer resources can sometimes be prohibitively expensive resulting in substantial computational demands. As far as probabilistic methods are concerned, they represent a useful tool, especially for planning studies, but, as evidenced by the many discussions reported in literature, they could reveal some shortcomings principally arising from: (1) the non-normal distribution and the statistical dependence of the input data; and (2) the difficulty arising in accurately identifying probability distributions for some input data, such as the power generated by wind or photovoltaic generators. All these could result in time consuming computations with several limitations in practical applications especially in power flow analysis of complex power networks. In order to try and overcome some of these limitations, obtaining thereby comprehensive power flow solution tolerance analysis at adequate computational costs, self validated computation could play a crucial role. Armed with such a vision, this chapter will analyze two advanced techniques for power flow analysis in the presence of data uncertainty namely the boundary power flow and the affine arithmetic power flow.

24 POWER TRANSMISSION AND DISTRIBUTION↗