Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “memory”

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 613 records · Page 34

MetaStrider: Architectures for Scalable Memory-centric Reduction of Sparse Data Streams

Reduction is an operation performed on the values of two or more key-value pairs that share the same key. Reduction of sparse data streams finds application in a wide variety of domains such as data and graph analytics, cybersecurity, machine learning, and HPC applications. However, these applications exhibit low locality of reference, rendering traditional architectures and data representations inefficient. This article presents MetaStrider, a significant algorithmic and architectural enhancement to the state-of-the-art, SuperStrider. Furthermore, these enhancements enable a variety of parallel, memory-centric architectures that we propose, resulting in demonstrated performance that scales near-linearly with available memory-level parallelism.

97 MATHEMATICS AND COMPUTING↗

Filament-Free Bulk Resistive Memory Enables Deterministic Analogue Switching

Digital computing is nearing its physical limits as computing needs and energy consumption rapidly increase. Analogue-memory-based neuromorphic computing can be orders of magnitude more energy efficient at data-intensive tasks like deep neural networks, but has been limited by the inaccurate and unpredictable switching of analogue resistive memory. Filamentary resistive random access memory (RRAM) suffers from stochastic switching due to the random kinetic motion of discrete defects in the nanometer-sized filament. Here, this stochasticity is overcome by incorporating a solid electrolyte interlayer, in this case, yttria-stabilized zirconia (YSZ), toward eliminating filaments. Filament-free, bulk-RRAM cells instead store analogue states using the bulk point defect concentration, yielding predictable switching because the statistical ensemble behavior of oxygen vacancy defects is deterministic even when individual defects are stochastic. Both experiments and modeling show bulk-RRAM devices using TiO2-X switching layers and YSZ electrolytes yield deterministic and linear analogue switching for efficient inference and training. Bulk-RRAM solves many outstanding issues with memristor unpredictability that have inhibited commercialization, and can, therefore, enable unprecedented new applications for energy-efficient neuromorphic computing. Beyond RRAM, this work shows how harnessing bulk point defects in ionic materials can be used to engineer deterministic nanoelectronic materials and devices.

36 MATERIALS SCIENCE↗

A fast, memory efficient and robust sparse preconditioner based on a multifrontal approach with applications to finite‐element matrices

Summary In this article, we introduce a fast, memory efficient and robust sparse preconditioner that is based on a direct factorization scheme for sparse matrices arising from the finite‐element discretization of elliptic partial differential equations. We use a fast (but approximate) multifrontal approach as a preconditioner and use an iterative scheme to achieve a desired accuracy. This approach combines the advantages of direct and iterative schemes to arrive at a fast, robust, and accurate preconditioner. We will show that this approach is faster (∼2×) and more memory efficient (∼2–3×) than a conventional direct multifrontal approach. Furthermore, we will demonstrate that this preconditioner is both faster and more effective than other preconditioners such as the incomplete LU preconditioner. Specific speedups depend on the matrix size and improve as the size of the matrix increases. The preconditioner can be applied to both structured and unstructured meshes in a similar manner. We build on our previous work and utilize the fact that dense frontal and update matrices, in the multifrontal algorithm, can be represented as hierarchically off‐diagonal low‐rank matrices. Using this idea, we replace all large dense matrix operations in the multifrontal elimination process with O ( N ) hierarchically off‐diagonal low‐rank operations to arrive at a faster and more memory efficient factorization scheme. We then use this direct factorization method at low accuracies as a preconditioner and apply it to various real‐life engineering test cases. Copyright © 2016 John Wiley & Sons, Ltd.

Aminfar, AmirHossein↗

Limited‐memory adaptive snapshot selection for proper orthogonal decomposition

Summary Reduced order models are useful for accelerating simulations in many‐query contexts, such as optimization, uncertainty quantification, and sensitivity analysis. However, offline training of reduced order models (ROMs) can have prohibitively expensive memory and floating‐point operation costs in high‐performance computing applications, where memory per core is limited. To overcome this limitation for proper orthogonal decomposition, we propose a novel adaptive selection method for snapshots in time that limits offline training costs by selecting snapshots according an error control mechanism similar to that found in adaptive time‐stepping ordinary differential equation solvers. The error estimator used in this work is related to theory bounding the approximation error in time of proper orthogonal decomposition‐based ROMs, and memory usage is minimized by computing the singular value decomposition using a single‐pass incremental algorithm. Results for a viscous Burgers' test problem demonstrate convergence in the limit as the algorithm error tolerances go to zero; in this limit, the full‐order model is recovered to within discretization error. A parallel version of the resulting method can be used on supercomputers to generate proper orthogonal decomposition‐based ROMs, or as a subroutine within hyperreduction algorithms that require taking snapshots in time, or within greedy algorithms for sampling parameter space. Copyright © 2016 John Wiley & Sons, Ltd.

Oxberry, Geoffrey M.↗

MAPredict: Static Analysis Driven Memory Access Prediction Framework for Modern CPUs

Application memory access patterns are crucial in deciding how much traffic is served by the cache and forwarded to the dynamic random-access memory (DRAM). However, predicting such memory traffic is difficult because of the interplay of prefetchers, compilers, parallel execution, and innovations in manufacturer-specific micro-architectures. This research introduced MAPredict, a static analysis-driven framework that addresses these challenges to predict last-level cache (LLC)-DRAM traffic. By exploring and analyzing the behavior of modern Intel processors, MAPredict formulates cache-aware analytical models. MAPredict invokes these models to predict LLC-DRAM traffic by combining the application model, machine model, and user-provided hints to capture dynamic information. MAPredict successfully predicts LLC-DRAM traffic for different regular access patterns and provides the means to combine static and empirical observations for irregular access patterns. Evaluating 130 workloads from six applications on recent Intel micro-architectures, MAPredict yielded an average accuracy of 99% for streaming, 91% for strided, and 92% for stencil patterns. By coupling static and empirical methods, up to 97% average accuracy was obtained for random access patterns on different micro-architectures.

Monil, M. A. H.↗

Analyzing the Energy and Power Consumption of Remote Memory Accesses in the OpenSHMEM Model

PGAS models like OpenSHMEM provide interfaces to explicitly initiate one-sided remote memory accesses among processes. In addition, the model also provides synchronizing barriers to ensure a consistent view of the distributed memory at different phases of an application. The incorrect use of such interfaces affects the scalability achievable while using a parallel programming model. This study aims at understanding the effects of these constructs on the energy and power consumption behavior of OpenSHMEM applications. Our experiments show that cost incurred in terms of the total energy and power consumed depends on multiple factors across the software and hardware stack. We conclude that there is a significant impact on the power consumed by the CPU and DRAM due to multiple factors including the design of the data transfer patterns within an application, the design of the communication protocols within a middleware, the architectural constraints laid by the interconnect solutions, and also the levels of memory hierarchy within a compute node. This work motivates treating energy and power consumption as important factors while designing compute solutions for current and future distributed systems.

Jana, Siddhartha↗

Multilayered microstructures with shape memory effects for vertical deployment

This paper presents a fabrication and characterization of multilayered microstructures with shape memory effects enabling large vertical deployment under electro-thermal actuation. Our previous research demonstrated vertical deployment of such microstructures by the effect of thermal mismatch. Development of equiatomic NiTi layers in the multilayered microstructure is investigated further for shape memory effects. Multilayered microstructures are built by sputtered NiTi layers and a lift-off process. Negative photoresist ma-N1420 enables clean lift-off of 500 nm thick NiTi layers by forming a significant undercut profile after development. The parametric study on co-sputter powers for Ti and Ni 50 Ti 50 targets suggests that 100 W RF on Ti target and 200 W DC on Ni 50 Ti 50 target can deposit Ni 49.62 Ti 50.38 layers. X-Ray Diffraction (XRD) and Atomic Force Microscopy (AFM) were used to study the crystal structures and surface topography of NiTi layers. XRD results of post-annealed Ni 49.62 Ti 50.38 layers show coexistence of austenite and martensitic phases at room temperature, suggesting that the transformation temperature of such NiTi layers should be approximate 20 °C. The surface topography of Ni 49.62 Ti 50.38 layers reveals substantial increase of surface roughness at ambient conditions after the annealing. Experimental verification of the multilayered microstructure for vertical deployment was carried out by Signatone Probe Station and Dual Scanning Electron Microscope/Focused Ion Beam (SEM/FIB) system. Finally, a vertical deployment of the two-dimensional (2D) multilayered microstructures for three-dimensional (3D) can be detected by applying a constant voltage of 0.04 V, and the expected 3D deployment displacement is enlarged from 2 μm to 10 μm by introducing the shape memory effect.

42 ENGINEERING↗

Memory-efficient nonsmooth dynamic optimization using adaptive randomized compression

Dynamic optimization problems arise in many applications including flow control, full waveform inversion, and medical imaging. These problems are plagued by significant computational challenges. One such challenge — and the focus of this work — is the memory limitation induced by the size of the underlying dynamical system. In particular, the entire dynamic trajectory is required for derivative computation and therefore must be stored or recomputed using, e.g., checkpointing. Although recent work demonstrated the use of adaptive randomized sketching to overcome the memory challenge, that work only applies to smooth unconstrained problems, prohibiting its use for nonsmooth regularized and constrained problems. The inclusion of nonsmooth regularizers and constraints is critical as they often arise in an attempt to preserve certain physical properties or to promote sparsity. To solve these problems, we introduce a trust-region algorithm for minimizing the sum of a smooth nonconvex function and a nonsmooth convex function that leverages randomized sketching to compress the dynamical system trajectories and adaptively adjust the sketch rank to satisfy a gradient inexactness condition. We prove convergence of this algorithm and demonstrate that it achieves substantial memory reduction on three discretized PDE-constrained optimization applications.

97 MATHEMATICS AND COMPUTING↗

Designing shape-memory-like microstructures in intercalation materials

During the reversible insertion of ions, lattices in intercalation materials undergo structural transformations. These lattice transformations generate misfit strains and volume changes that, in turn, contribute to the structural decay of intercalation materials and limit their reversible cycling. In this paper, we draw on insights from shape-memory alloys, another class of phase transformation materials, that also undergo large lattice transformations but do so with negligible macroscopic volume changes and internal stresses. We develop a theoretical framework to predict structural transformations in intercalation compounds and identify crystallographic design rules necessary for forming shape-memory-like microstructures in intercalation materials. We use our approach to systematically screen open-source structural databases comprising $n$ > 5000 pairs of intercalation compounds. We identify candidate compounds, such as Li x Mn 2 O 4 (Spinel), Li x Ti 2 (PO 4 ) 3 (NASICON), that approximately satisfy the crystallographic design rules and can be precisely doped to form shape-memory-like microstructures. Throughout, we compare our analytical results with experimental measurements of intercalation compounds. We find a direct correlation between structural transformations, microstructures, and increased capacity retention in these materials. These results, more generally, show that crystallographic designing of intercalation materials could be a novel route to discovering compounds that do not decay with continuous usage.

36 MATERIALS SCIENCE↗

Resistance switching memory in perovskite oxides

The resistance switching behavior has recently attracted great attentions for its application as resistive random access memories (RRAMs) due to a variety of advantages such as simple structure, high-density, high-speed and low-power. As a leading storage media, the transition metal perovskite oxide owns the strong correlation of electrons and the stable crystal structure, which brings out multifunctionality such as ferroelectric, multiferroic, superconductor, and colossal magnetoresistance/electroresistance effect, etc. The existence of rich electronic phases, metal–insulator transition and the nonstoichiometric oxygen in perovskite oxide provides good platforms to insight into the resistive switching mechanisms. In this review, we first introduce the general characteristics of the resistance switching effects, the operation methods and the storage media. Then, the experimental evidences of conductive filaments, the transport and switching mechanisms, and the memory performances and enhancing methods of perovskite oxide based filamentary RRAM cells have been summarized and discussed. Subsequently, the switching mechanisms and the performances of the uniform RRAM cells associating with the carrier trapping/detrapping and the ferroelectric polarization switching have been discussed. Finally, the advices and outlook for further investigating the resistance switching and enhancing the memory performances are given.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Repeated exposure of the developing rat brain to magnetic resonance imaging did not affect neurogenesis, cell death or memory function

Research highlights: {yields} The effect of MRI on the developing brain is a matter of debate. {yields} Repeated exposure to MRI did not affect neurogenesis. {yields} Memory function was not affected by repeated MRI during development. {yields} Neither late gestation nor young postnatal brains were affected by MRI. {yields} Repeated MRI did not cause cell death in the neurogenic region of the hippocampus. -- Abstract: The effect of magnetic fields on the brain is a matter of debate. The objective of this study was to investigate whether repeated exposure to strong magnetic fields, such as during magnetic resonance imaging (MRI), could elicit changes in the developing rat brain. Embryonic day 15 (E15) and postnatal day 14 (P14) rats were exposed to MRI using a 7.05 T MR system. The animals were anesthetized and exposed for 35 min per day for 4 successive days. Control animals were anesthetized but no MRI was performed. Body temperature was maintained at 37 {sup o}C. BrdU was injected after each session (50 mg/kg). One month later, cell proliferation, neurogenesis and astrogenesis in the dentate gyrus were evaluated, revealing no effects of MRI, neither in the E15, nor in the P14 group. DNA damage in the dentate gyrus in the P14 group was evaluated on P18, 1 day after the last session, using TUNEL staining. There was no difference in the number of TUNEL-positive cells after MRI compared with controls, neither in mature neurons, nor in newborn progenitors (BrdU/TUNEL double-labeled cells). Novel object recognition was performed to assess memory function 1 month after MRI. There was no difference in the recognition index observed after MRI compared with the control rats, neither for the E15, nor for the P14 group. In conclusion, repeated exposure to MRI did not appear to affect neurogenesis, cell death or memory function in rats, neither in late gestation (E15-E18) nor in young postnatal (P14-P17) rats.

60 APPLIED LIFE SCIENCES↗

Learning and Memory Following Conformal Radiation Therapy for Pediatric Craniopharyngioma and Low-Grade Glioma

Purpose: The primary objective of this study was to examine whether children with low-grade glioma (LGG) or craniopharyngioma had impaired learning and memory after conformal radiation therapy (CRT). A secondary objective was to determine whether children who received chemotherapy before CRT, a treatment often used to delay radiation therapy in younger children with LGG, received any protective benefit with respect to learning. Methods and Materials: Learning and memory in 57 children with LGG and 44 children with craniopharyngioma were assessed with the California Verbal Learning Test-Children's Version and the Visual-Auditory Learning tests. Learning measures were administered before CRT, 6 months later, and then yearly for a total of 5 years. Results: No decline in learning scores after CRT was observed when patients were grouped by diagnosis. For children with LGG, chemotherapy before CRT did not provide a protective effect on learning. Multiple regression analyses, which accounted for age and tumor volume and location, found that children treated with chemotherapy before CRT were at greater risk of decline on learning measures than those treated with CRT alone. Variables predictive of learning and memory decline included hydrocephalus, shunt insertion, younger age at time of treatment, female gender, and pre-CRT chemotherapy. Conclusions: This study did not reveal any impairment or decline in learning after CRT in overall aggregate learning scores. However, several important variables were found to have a significant effect on neurocognitive outcome. Specifically, chemotherapy before CRT was predictive of worse outcome on verbal learning in LGG patients. In addition, hydrocephalus and shunt insertion in craniopharyngioma were found to be predictive of worse neurocognitive outcome, suggesting a more aggressive natural history for those patients.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Stochastic goal-oriented error estimation with memory

We propose a stochastic dual-weighted error estimator for the viscous shallow-water equation with boundaries. For this purpose, previous work on memory-less stochastic dual-weighted error estimation is extended by incorporating memory effects. The memory is introduced by describing the local truncation error as a sum of time-correlated random variables. The random variables itself represent the temporal fluctuations in local truncation errors and are estimated from high-resolution information at near-initial times. The resulting error estimator is evaluated experimentally in two classical ocean-type experiments, the Munk gyre and the flow around an island. In these experiments, the stochastic process is adapted locally to the respective dynamical flow regime. Our stochastic dual-weighted error estimator is shown to provide meaningful error bounds for a range of physically relevant goals. We prove, as well as show numerically, that our approach can be interpreted as a linearized stochastic-physics ensemble.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Memory-optimized shift operator alternating direction implicit finite difference time domain method for plasma

Through introducing the alternating direction implicit (ADI) technique and the memory-optimized algorithm to the shift operator (SO) finite difference time domain (FDTD) method, the memory-optimized SO-ADI FDTD for nonmagnetized collisional plasma is proposed and the corresponding formulae of the proposed method for programming are deduced. In order to further the computational efficiency, the iteration method rather than Gauss elimination method is employed to solve the equation set in the derivation of the formulae. Complicated transformations and convolutions are avoided in the proposed method compared with the Z transforms (ZT) ADI FDTD method and the piecewise linear JE recursive convolution (PLJERC) ADI FDTD method. The numerical dispersion of the SO-ADI FDTD method with different plasma frequencies and electron collision frequencies is analyzed and the appropriate ratio of grid size to the minimum wavelength is given. The accuracy of the proposed method is validated by the reflection coefficient test on a nonmagnetized collisional plasma sheet. The testing results show that the proposed method is advantageous for improving computational efficiency and saving computer memory. The reflection coefficient of a perfect electric conductor (PEC) sheet covered by multilayer plasma and the RCS of the objects coated by plasma are calculated by the proposed method and the simulation results are analyzed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Growth and self-assembly of BaTiO{sub 3} nanocubes for resistive switching memory cells

In this work, the self-assembled BaTiO{sub 3} nanocubes based resistive switching memory capacitors are fabricated with hydrothermal and drop-coating approaches. The device exhibits excellent bipolar resistance switching characteristics with ON/OFF ratio of 58–70, better reliability and stability over various polycrystalline BaTiO{sub 3} nanostructures. It is believed that the inter cube junctions is responsible for such a switching behaviour and it can be described by the filament model. The effect of film thickness on switching ratio (ON/OFF) was also investigated in details. - Graphical abstract: This work describes a novel resistive switching memory cell based on self-assembled BaTiO{sub 3} nanocubes. - Highlights: • BaTiO{sub 3} nanocubes were prepared by one step facile hydrothermal method. • Self-assembled BaTiO{sub 3} nanocubes thin films were obtained by drop-coating approach. • The BaTiO{sub 3} nanocubes show excellent resistive switching properties for memory applications.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Towards scaling community detection on distributed-memory heterogeneous systems

Distributed multi-GPU systems pose significant challenges and opportunities for efficient execution of parallel applications. Graph algorithms are generally characterized by irregular memory accesses, low computation to communication ratios, and load balancing problems that are especially hard to address on multi-GPU systems. Graph community detection is an important problem in the emerging domain of graph analytics with numerous applications. In this paper, we present our ongoing work on distributed-memory multi-GPU implementation for graph community detection. Our work parallelizes the widely used (albeit serial) Louvain method on distributed multi-GPU platforms. Supported by an extensive set of experiments on a multi-GPU enabled supercomputer (OLCF Summit) and a single compute node (Nvidia DGX-2®), we demonstrate competitive performance to existing distributed-memory CPU-based implementation, and up to 6.5 better results than Nvidia RAPIDS® CUGRAPH. To the best of our knowledge, this work represents the first effort for community detection on distributed multi-GPU systems. Our approach and related findings can be extended to numerous other iterative graph algorithms on multi-GPU systems.

97 MATHEMATICS AND COMPUTING↗

Biochemical modeling of microbial memory effects and catabolite repression on soil organic carbon compounds

Here, we report that microbial decomposition of Soil Organic Matter (SOM) is largely controlled by environmental and edaphic factors such as temperature, pH, and moisture. However, microbial metabolism is controlled by catabolite repression, which leads microbes to grow on preferred nutrient and energy sources first. In particular, Catabolite Repression for Carbon (CR-C) defines the hierarchical preference of bacteria for particular C sources. This control depends on the presence of signal molecules conferring bacteria a memory for recent growth conditions on less preferred C sources. The combined effect of catabolite repression and microbial memory (called here Memory-Associated Catabolite Repression for Carbon, MACR-C) has not yet been investigated in detail. First, we use observations and a numerical model to test the hypothesis that MACR-C explains substrate preferential consumption in a simple, 2-C substrate system, whereas Michaelis-Menten-Monod kinetics of competitive substrate consumption, non-competitive inhibition, or their combination, do not. Next, we carry out numerical analyses to explore the sensitivity of (1) estimated parameters to experimental observations and (2) model structure to steady-state substrate concentration under pulse or continuous substrate application. These results show that MACR-C substantially affected substrate consumption and microbial readiness to switch between C sources.

54 ENVIRONMENTAL SCIENCES↗

{gamma}-aminobutyric acid{sub A} (GABA{sub A}) receptor regulates ERK1/2 phosphorylation in rat hippocampus in high doses of Methyl Tert-Butyl Ether (MTBE)-induced impairment of spatial memory

Experimental and occupational exposure to Methyl Tert-Butyl Ether (MTBE) has been reported to induce neurotoxicological and neurobehavioral effects, such as headache, nausea, dizziness, and disorientation, etc. However, the molecular mechanisms involved in MTBE-induced neurotoxicity are still not well understood. In the present study, we investigated the effects of MTBE on spatial memory and the expression and function of GABA{sub A} receptor in the hippocampus. Our results demonstrated that intraventricular injection of MTBE impaired the performance of the rats in a Morris water maze task, and significantly increased the expression of GABA{sub A} receptor {alpha}1 subunit in the hippocampus. The phosphorylation of ERK1/2 decreased after the MTBE injection. Furthermore, the decreased ability of learning and the reduction of phosphorylated ERK1/2 level of the MTBE-treated rats was partly reversed by bicuculline injected 30 min before the training. These results suggested that MTBE exposure could result in impaired spatial memory. GABA{sub A} receptor may play an important role in the MTBE-induced impairment of learning and memory by regulating the phosphorylation of ERK in the hippocampus.

60 APPLIED LIFE SCIENCES↗