Search NASA⌕ Search

SEARCH · Search NASA

Results for “complex computing”

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

On Bellman’s and Knuth’s Problems and their Generalizations

Various generalizations of the classical problem of the fastest raising to a power (or the so-called problem on addition chains) are studied in the asymptotic sense. Under weak restrictions, we demonstrate asymptotically tight solutions of the two best known generalizations, namely, Bellman’s problem on the computational complexity (on the minimal number of multiplication operations) of a normed monomial of several variables and Knuth’s problem on the computational complexity of a power system of one variable. We also briefly review some results on the computational complexity for three problems, namely, the computation of p-element systems of normed monomials in q variables, additive computations for systems of p integer linear forms over q variables, and the computation of p-element systems of the free Abelian group with q generators.

97 MATHEMATICS AND COMPUTING↗

The Tiny Triplet Finder as a Versatile Track Segment Seeding Engine for Trigger Systems

In high energy physics experiment trigger systems, track segment seeding is a resource consuming function and the primary reason is the computing complexity of the segment finding process. As the Moore's Law is reaching its physical limit, reducing computing complexity should be carefully considered, rather than keep piling up silicon resources. The Tiny Triplet Finder is a scheme that reduces the computing complexity of the segment seeding. As a proof of concept, a 3D track segment seeding engine core based on the Tiny Triplet Finder has been implemented and tested in a low-cost FPGA device. The seeding engine is designed to preselect and group hits (stubs) from detector layers to feed subsequent track fitting stage. The seeding engine consists of a Hough transform space for r-z view and a Tiny Triplet Finder for r-phi view to implement 3D constraints. The seeding engine is organized as a pipeline so that each hit is processed in a single clock cycle. Taking advantage of the register-like storage block scheme which enables effectively resetting of a block RAM within a single clock cycle, clearing or refreshing the seeding engine takes only one clock cycles between two events. The Tiny Triplet Finder is also a generic coincidence finding scheme that can be used for many tasks. As a versatility demonstration, track segment finding performances for two distinctive detector geometries are tested in our seeding engine. In a collider barrel-layer geometry, the fake segment rates are studied for 3D (i.e., both r-phi and r-z views) and 2D (i.e., r-phi or r-z view only) configurations for high hit multiplicity events (>4000 hits/layer in the barrel region). Another detector geometry contains strip plane layers with timing information. The numbers of coincidences, both real or fake, with or without timing ("3D" or "2D") information at various hit multiplicities are studied.

43 PARTICLE ACCELERATORS↗

Efficient and Robust Jet Tagging at the LHC with Knowledge Distillation

The challenging environment of real-time data processing systems at the Large Hadron Collider (LHC) strictly limits the computational complexity of algorithms that can be deployed. For deep learning models, this implies that only models with low computational complexity that have weak inductive bias are feasible. To address this issue, we utilize knowledge distillation to leverage both the performance of large models and the reduced computational complexity of small ones. In this paper, we present an implementation of knowledge distillation, demonstrating an overall boost in the student models' performance for the task of classifying jets at the LHC. Furthermore, by using a teacher model with a strong inductive bias of Lorentz symmetry, we show that we can induce the same inductive bias in the student model which leads to better robustness against arbitrary Lorentz boost.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Trajectory prediction dimensionality reduction for low-cost connected automated vehicle systems

Here, to facilitate low-cost connected automated vehicle (CAV) system development, this study proposes two interpretable dimensionality reduction techniques in vehicle trajectory prediction, i.e., the piecewise Taylor series approximation (PTA) and the piecewise Fourier series approximation (PFA), to lower computation complexity, reduce device investment, and decrease computation energy consumption. Two benchmarks are developed, the long short-term memory (LSTM)-based model without dimensionality reduction and the LSTM-based model with encoder-decoder (a widely used dimensionality reduction technique). Results show that the four predictions have similar accuracy, and the training time (proportional to computation energy consumption) of models with dimensionality reduction techniques is greatly reduced. The reduction is even more significant when PTA/PFA is used. Sensitivity analysis advises PFA/PTA parameter selections to reduce computation complexity without significant loss of prediction accuracy. Further, the robustness of the LSTM PTA/PFA is proven by the investigation of data noises.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Where are the parallel algorithms?

Four paradigms that can be useful in developing parallel algorithms are discussed. These include computational complexity analysis, changing the order of computation, asynchronous computation, and divide and conquer. Each is illustrated with an example from scientific computation, and it is shown that computational complexity must be used with great care or an inefficient algorithm may be selected.

Voigt, R. G.↗

Mathematical model and software complex for computer simulation of field emission electron sources

The software complex developed in MATLAB allows modelling of function of diode and triode structures based on field emission electron sources with complex sub-micron geometry, their volt-ampere characteristics, calculating distribution of electric field for educational and research needs. The goal of this paper is describing the physical-mathematical model, calculation methods and algorithms the software complex is based on, demonstrating the principles of its function and showing results of its work. For getting to know the complex, a demo version with graphical user interface is presented.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A dynamically bi-orthogonal method for time-dependent stochastic partial differential equations II: Adaptivity and generalizations

This is part II of our paper in which we propose and develop a dynamically bi-orthogonal method (DyBO) to study a class of time-dependent stochastic partial differential equations (SPDEs) whose solutions enjoy a low-dimensional structure. In part I of our paper [9], we derived the DyBO formulation and proposed numerical algorithms based on this formulation. Some important theoretical results regarding consistency and bi-orthogonality preservation were also established in the first part along with a range of numerical examples to illustrate the effectiveness of the DyBO method. In this paper, we focus on the computational complexity analysis and develop an effective adaptivity strategy to add or remove modes dynamically. Our complexity analysis shows that the ratio of computational complexities between the DyBO method and a generalized polynomial chaos method (gPC) is roughly of order O((m/N{sub p}){sup 3}) for a quadratic nonlinear SPDE, where m is the number of mode pairs used in the DyBO method and N{sub p} is the number of elements in the polynomial basis in gPC. The effective dimensions of the stochastic solutions have been found to be small in many applications, so we can expect m is much smaller than N{sub p} and computational savings of our DyBO method against gPC are dramatic. The adaptive strategy plays an essential role for the DyBO method to be effective in solving some challenging problems. Another important contribution of this paper is the generalization of the DyBO formulation for a system of time-dependent SPDEs. Several numerical examples are provided to demonstrate the effectiveness of our method, including the Navier–Stokes equations and the Boussinesq approximation with Brownian forcing.

97 MATHEMATICS AND COMPUTING↗

A molecular approach to understanding complex systems: computational statistical mechanics using state-of-the-art algorithms on terascale computational platforms

Presented herein, is our recent progress towards a molecular picture of reactions in complex fluids and there interfaces. In particular, we highlight the important statistical mechanical methodology as well as the electronic structure methods and algorithms that are used to tackle important unresolved questions in the chemical sciences. Moreover, we demonstrate that though the simulation protocol outlined, in conjunction with current and next generation computer hardware, that the elucidation of complex chemical problems in the condensed phase are within reach. This work was supported by the U.S. Department of Energy's (DOE) Office of Basic Energy Sciences, Chemical Sciences program. The Pacific Northwest National Laboratory is operated by Battelle for DOE.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A unifying framework for rigid multibody dynamics and serial and parallel computational issues

A unifying framework for various formulations of the dynamics of open-chain rigid multibody systems is discussed. Their suitability for serial and parallel processing is assessed. The framework is based on the derivation of intrinsic, i.e., coordinate-free, equations of the algorithms which provides a suitable abstraction and permits a distinction to be made between the computational redundancy in the intrinsic and extrinsic equations. A set of spatial notation is used which allows the derivation of the various algorithms in a common setting and thus clarifies the relationships among them. The three classes of algorithms viz., O(n), O(n exp 2) and O(n exp 3) or the solution of the dynamics problem are investigated. Researchers begin with the derivation of O(n exp 3) algorithms based on the explicit computation of the mass matrix and it provides insight into the underlying basis of the O(n) algorithms. From a computational perspective, the optimal choice of a coordinate frame for the projection of the intrinsic equations is discussed and the serial computational complexity of the different algorithms is evaluated. The three classes of algorithms are also analyzed for suitability for parallel processing. It is shown that the problem belongs to the class of N C and the time and processor bounds are of O(log2/2(n)) and O(n exp 4), respectively. However, the algorithm that achieves the above bounds is not stable. Researchers show that the fastest stable parallel algorithm achieves a computational complexity of O(n) with O(n exp 4), respectively. However, the algorithm that achieves the above bounds is not stable. Researchers show that the fastest stable parallel algorithm achieves a computational complexity of O(n) with O(n exp 2) processors, and results from the parallelization of the O(n exp 3) serial algorithm.

Fijany, Amir↗

Leveraging dendritic complexity for neuromorphic computing

Abstract Beyond-von Neumann computing approaches are necessary to sustain the growth of microelectronics and the increasing appetite for artificial intelligence/machine learning algorithms. Neuromorphic computing is an emerging paradigm that takes inspiration from the brain to provide a path forward to improve the computational efficiency and computational density of next-generation computing architectures. In nature, we observe brains performing complex computations with a much smaller energy footprint than conventional computing approaches. Current neuromorphic systems are focused primarily on scalability, namely, increasing the number of computational units (neurons) and connections between units (synapses). However, for brain-like cognition and efficiency in next-generation computing hardware, we need increased complexity in function, as well as improved connection density for scalability. Here, we present our work that aims to incorporate dendrites for ‘compute-on-wire’ in neuromorphic architectures to increase the computational complexity (e.g. number of programmable parameters, nonlinear dynamics) as well as computational efficiency (energy/compute) of artificial neural networks (ANNs). We do this by showcasing neuromorphic dendrite elements that can be leveraged for various applications. We will present examples of neuroscience-inspired direction-selective circuits and an ANN with active dendrites leveraging shunting inhibition. We also demonstrate the benefits of using dendrites in deep neural networks. To conclude, we discuss how we can utilize emerging hardware devices in these systems and design next-generation neuromorphic architectures with dendrites.

Cardwell, Suma G. (ORCID:0000000226575545)↗

Bridge-Mediated Metal-to-Metal Electron and Hole Transfer in a Supermolecular Dinuclear Complex: A Computational Study Using Quantum Electron–Nuclear Dynamics

Bimetallic electron donor–acceptor complexes can facilitate electron and energy transfer with excellent structural control through synthetic design. In this work, we investigate the photochemical dynamics in a Ru–Cu bimetallic complex after photoexcitation of the Ru-centered charge transfer state. The physical underpinnings of the metal-to-metal directional charge transfer process are unraveled via analyses of the quantum electronic dynamics and electron–nuclear trajectories. Here, the effects of molecular vibrations in the photoexcited state on the charge transfer processes are also analyzed.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

FLAME: A platform for high performance computing of complex systems, applied for three case studies

FLAME allows complex models to be automatically parallelised on High Performance Computing (HPC) grids enabling large number of agents to be simulated over short periods of time. Modellers are hindered by complexities of porting models on parallel platforms and time taken to run large simulations on a single machine, which FLAME overcomes. Three case studies from different disciplines were modelled using FLAME, and are presented along with their performance results on a grid.

97 MATHEMATICS AND COMPUTING↗

Building a Foundation for Structure-Based Cellulosome Design for Cellulosic Ethanol: Insight into Cohesin-Dockerin Complexation from Computer Simulation

The organization and assembly of the cellulosome, an extracellular multienzyme complex produced by anaerobic bacteria, is mediated by the high-affinity interaction of cohesin domains from scaffolding proteins with dockerins of cellulosomal enzymes. We have performed molecular dynamics simulations and free energy calculations on both the wild type (WT) and D39N mutant of the C. thermocellum Type I cohesin-dockerin complex in aqueous solution. The D39N mutation has been experimentally demonstrated to disrupt cohesin-dockerin binding. The present MD simulations indicate that the substitution triggers significant protein flexibility and causes a major change of the hydrogen-bonding network in the recognition strips - the conserved loop regions previously proposed to be involved in binding - through electrostatic and salt-bridge interactions between {beta}-strands 3 and 5 of the cohesin and {alpha}-helix 3 of the dockerin. The mutation-induced subtle disturbance in the local hydrogen-bond network is accompanied by conformational rearrangements of the protein side chains and bound water molecules. Additional free energy perturbation calculations of the D39N mutation provide differences in the cohesin-dockerin binding energy, thus offering a direct, quantitative comparison with experiments. The underlying molecular mechanism of cohesin-dockerin complexation is further investigated through the free energy profile, that is, potential of mean force (PMF) calculations of WT cohesin-dockerin complex. The PMF shows a high-free energy barrier against the dissociation and reveals a stepwise pattern involving both the central {beta}-sheet interface and its adjacent solvent-exposed loop/turn regions clustered at both ends of the {beta}-barrel structure.

adaptive biasing force↗

Scalable Hyper-parameter Estimation for Gaussian Process Based Time Series Analysis

Gaussian process (GP) is increasingly becoming popular as a kernel machine learning tool for non-parametric data analysis. Recently, GP has been applied to model non-linear dependencies in time series data. GP based analysis can be used to solve problems of time series prediction, forecasting, missing data imputation, change point detection, anomaly detection, etc. But the use of GP to handle massive scientific time series data sets has been limited, owing to its expensive computational complexity. The primary bottleneck is the handling of the covariance matrix whose size is quadratic in the length of the time series. In this paper we propose a scalable method that exploit the special structure of the covariance matrix for hyper-parameter estimation in GP based learning. The proposed method allows estimation of hyper parameters associated with GP in quadratic time, which is an order of magnitude improvement over standard methods with cubic complexity. Moreover, the proposed method does not require explicit computation of the covariance matrix and hence has memory requirement linear to the length of the time series as opposed to the quadratic memory requirement of standard methods. To further improve the computational complexity of the proposed method, we provide a parallel version to concurrently estimate the log likelihood for a set of time series which is the key step in the hyper-parameter estimation. Performance results on a multi-core system show that our proposed method provides significant speedups as high as 1000, even when running in serial mode, while maintaining a small memory footprint. The parallel version exploits the natural parallelization potential of the serial algorithm and is shown to perform significantly better than the serial faster algorithm, with speedups as high as 10.

Chandola, Varun↗

On the recognition of complex structures: Computer software using artificial intelligence applied to pattern recognition

An approach to simultaneous interpretation of objects in complex structures so as to maximize a combined utility function is presented. Results of the application of a computer software system to assign meaning to regions in a segmented image based on the principles described in this paper and on a special interactive sequential classification learning system, which is referenced, are demonstrated.

Yakimovsky, Y.↗

Capturing provenance as a diagnostic tool for workflow performance evaluation and optimization

In extreme-scale computing environments such as the DOE Leadership Computing Facilities scientific workflows are routinely used to coordinate software processes for the execution of complex, computational applications that perform in-silico experiments. Monitoring the performance of workflows without also simultaneously tracking provenance is not sufficient to understand variations between runs, configurations, versions of a code, and between changes in an implemented stack, and systems, i.e. the variability of performance metrics data in their historical context. We take a provenance-based approach and demonstrate that provenance is useful as a tool for evaluating and optimizing workflow performance in extreme-scale HPC environments. We present Chimbuko, a framework for the analysis and visualization of the provenance of performance. Chimbuko implements a method for the evaluation of workflow performance from multiple components that enables the exploration of performance metrics data at scale.

97 MATHEMATICS AND COMPUTING↗