Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “MEMORY DEVICES”

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 1,675 records · Page 93

Considerations on the Use of Custom Accelerators for Big Data Analytics

Accelerators, including Graphic Processing Units (GPUs) for gen- eral purpose computation, many-core designs with wide vector units (e.g., Intel Phi), have become a common component of many high performance clusters. The appearance of more stable and reliable tools tools that can automatically convert code written in high-level specifications with annotations (such as C or C++) to hardware de- scription languages (High-Level Synthesis - HLS), is also setting the stage for a broader use of reconfigurable devices (e.g., Field Pro- grammable Gate Arrays - FPGAs) in high performance system for the implementation of custom accelerators, helped by the fact that new processors include advanced cache-coherent interconnects for these components. In this chapter, we briefly survey the status of the use of accelerators in high performance systems targeted at big data analytics applications. We argue that, although the progress in the use of accelerators for this class of applications has been sig- nificant, differently from scientific simulations there still are gaps to close. This is particularly true for the ”irregular” behaviors exhibited by no-SQL, graph databases. We focus our attention on the limits of HLS tools for data analytics and graph methods, and discuss a new architectural template that better fits the requirement of this class of applications. We validate the new architectural templates by mod- ifying the Graph Engine for Multithreaded System (GEMS) frame- work to support accelerators generated with such a methodology, and testing with queries coming from the Lehigh University Benchmark (LUBM). The architectural template enables better supporting the task and memory level parallelism present in graph methods by sup- porting a new control model and a enhanced memory interface. We show that out solution allows generating parallel accelerators, pro- viding speed ups with respect to conventional HLS flows. We finally draw conclusions and present a perspective on the use of reconfig- urable devices and Design Automation tools for data analytics.

Data Analytics↗

Perpendicular spin transfer torque magnetic random access memories with high spin torque efficiency and thermal stability for embedded applications (invited)

Magnetic random access memories based on the spin transfer torque phenomenon (STT-MRAMs) have become one of the leading candidates for next generation memory applications. Among the many attractive features of this technology are its potential for high speed and endurance, read signal margin, low power consumption, scalability, and non-volatility. In this paper, we discuss our recent results on perpendicular STT-MRAM stack designs that show STT efficiency higher than 5 k{sub B}T/μA, energy barriers higher than 100 k{sub B}T at room temperature for sub-40 nm diameter devices, and tunnel magnetoresistance higher than 150%. We use both single device data and results from 8 Mb array to demonstrate data retention sufficient for automotive applications. Moreover, we also demonstrate for the first time thermal stability up to 400 °C exceeding the requirement of Si CMOS back-end processing, thus opening the realm of non-volatile embedded memory to STT-MRAM technology.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Roofline Analysis in the Intel® Advisor to Deliver Optimized Performance for applications on Intel® Xeon Phi™ Processor

In this session we show, in two case studies, how the roofline feature of Intel Advisor has been utilized to optimize the performance of kernels of the XGC1 and PICSAR codes in preparation for Intel Knights Landing architecture. The impact of the implemented optimizations and the benefits of using the automatic roofline feature of Intel Advisor to study performance of large applications will be presented. This demonstrates an effective optimization strategy that has enabled these science applications to achieve up to 4.6 times speed-up and prepare for future exascale architectures. # Goal/Relevance of Session The roofline model [1,2] is a powerful tool for analyzing the performance of applications with respect to the theoretical peak achievable on a given computer architecture. It allows one to graphically represent the performance of an application in terms of operational intensity, i.e. the ratio of flops performed and bytes moved from memory in order to guide optimization efforts. Given the scale and complexity of modern science applications, it can often be a tedious task for the user to perform the analysis on the level of functions or loops to identify where performance gains can be made. With new Intel tools, it is now possible to automate this task, as well as base the estimates of peak performance on measurements rather than vendor specifications. The goal of this session is to demonstrate how the roofline feature of Intel Advisor can be used to balance memory vs. computation related optimization efforts and effectively identify performance bottlenecks. A series of typical optimization techniques: cache blocking, structure refactoring, data alignment, and vectorization illustrated by the kernel cases will be addressed. # Description of the codes ## XGC1 The XGC1 code [3] is a magnetic fusion Particle-In-Cell code that uses an unstructured mesh for its Poisson solver that allows it to accurately resolve the edge plasma of a magnetic fusion device. After recent optimizations to its collision kernel [4], most of the computing time is spent in the electron push (pushe) kernel, where these optimization efforts have been focused. The kernel code scaled well with MPI+OpenMP but had almost no automatic compiler vectorization, in part due to indirect memory addresses and in part due to low trip counts of low-level loops that would be candidates for vectorization. Particle blocking and sorting have been implemented to increase trip counts of low-level loops and improve memory locality, and OpenMP directives have been added to vectorize compute-intensive loops that were identified by Advisor. The optimizations have improved the performance of the pushe kernel 2x on Haswell processors and 1.7x on KNL. The KNL node-for-node performance has been brought to within 30% of a NERSC Cori phase I Haswell node and we expect to bridge this gap by reducing the memory footprint of compute intensive routines to improve cache reuse. ## PICSAR is a Fortran/Python high-performance Particle-In-Cell library targeting at MIC architectures first designed to be coupled with the PIC code WARP for the simulation of laser-matter interaction and particle accelerators. PICSAR also contains a FORTRAN stand-alone kernel for performance studies and benchmarks. A MPI domain decomposition is used between NUMA domains and a tile decomposition (cache-blocking) handled by OpenMP has been added for shared-memory parallelism and better cache management. The so-called current deposition and field gathering steps that compose the PIC time loop constitute major hotspots that have been rewritten to enable more efficient vectorization. Particle communications between tiles and MPI domain has been merged and parallelized. All considered, these improvements provide speedups of 3.1 for order 1 and 4.6 for order 3 interpolation shape factors on KNL configured in SNC4 quadrant flat mode. Performance is similar between a node of cori phase 1 and KNL at order 1 and better on KNL by a factor 1.6 at order 3 with the considered test case (homogeneous thermal plasma).

97 MATHEMATICS AND COMPUTING↗

On the Feasibility of 1T Ferroelectric FET Memory Array

To fully exploit the ferroelectric field effect transistor (FeFET) as compact embedded nonvolatile memory for various computing and storage applications, it is desirable to use a single FeFET (1T) as a unit cell and arrange the cells into an array. However, many write mechanisms for an 1T FeFET array reported in the literature are yet to be validated experimentally. In this work, we performed a comprehensive experimental characterization on the write operations in an 1T- NOR and 1T- AND array using n-channel bulk FeFETs. We discovered that: 1) the source/drain contact can only supply minority carriers (i.e., electrons) to the channel for polarization screening during the low- V TH state programming; 2) the body contact can not only supply majority carriers (i.e., holes) for efficient high- V TH state write, but also depletion charge for low- V TH state programming, though with lower efficiency; and 3) during the low/high- V TH programming, only the path that can supply negative/positive screening charges, respectively, need to respond, which necessitates the application of proper write biases on the corresponding terminal. Based on the understanding of these write mechanisms, we show the importance of localized body contact or column-wise body contact for the successful high- V TH state programming. We also show that the previously proposed C- AND write scheme fails to program the FeFET to the low- V TH state for our devices. Finally, we propose several write schemes for both 1T- AND and 1T- NOR arrays for various scenarios providing insights for choosing the appropriate write scheme, which will facilitate the adoption of 1T FeFET memory arrays for emerging applications.

FeFET↗

Nanoelectronics with proximitized materials

While materials design for many device applications usually relies on adding impurities, recent advances in scaling-down heterostructures with improved interfacial properties offer a different way to transform a large class of materials. A given material can be drastically changed by inheriting properties leaking from its neighboring regions, such as magnetism, superconductivity, or spin-orbit coupling. While these proximity effects often have a short range and are considered negligible, the situation is qualitatively different in atomically thin and two-dimensional materials where the extent of proximity effects can exceed their thickness. Consequently, proximitized materials have a potential to display novel properties and device opportunities, absent in any of the constituent region of the considered heterostructures. Such proximitized materials could provide platforms for a wide range of emerging applications: from seamless integration of memory and logic, to fault-tolerant topologically protected quantum computing.

36 MATERIALS SCIENCE↗

Influence of thermal agitation on the electric field induced precessional magnetization reversal with perpendicular easy axis

We investigated the influence of thermal agitation on the electric field induced precessional magnetization switching probability with perpendicular easy axis by solving the Fokker-Planck equation numerically with finite difference method. The calculated results show that the thermal agitation during the reversal process crucially influences the switching probability. The switching probability can be achieved is only determined by the thermal stability factor Δ of the free layer, it is independent on the device dimension, which is important for the high density device application. Ultra-low error rate down to the order of 10{sup −9} can be achieved for the device of thermal stability factor Δ of 40. Low damping factor α material should be used for the free layer for high reliability device applications. These results exhibit potential of electric field induced precessional magnetization switching with perpendicular easy axis for ultra-low power, high speed and high density magnetic random access memory (MRAM) applications.

36 MATERIALS SCIENCE↗

Roadmap for unconventional computing with nanotechnology

Abstract In the ‘Beyond Moore’s Law’ era, with increasing edge intelligence, domain-specific computing embracing unconventional approaches will become increasingly prevalent. At the same time, adopting a variety of nanotechnologies will offer benefits in energy cost, computational speed, reduced footprint, cyber resilience, and processing power. The time is ripe for a roadmap for unconventional computing with nanotechnologies to guide future research, and this collection aims to fill that need. The authors provide a comprehensive roadmap for neuromorphic computing using electron spins, memristive devices, two-dimensional nanomaterials, nanomagnets, and various dynamical systems. They also address other paradigms such as Ising machines, Bayesian inference engines, probabilistic computing with p-bits, processing in memory, quantum memories and algorithms, computing with skyrmions and spin waves, and brain-inspired computing for incremental learning and problem-solving in severely resource-constrained environments. These approaches have advantages over traditional Boolean computing based on von Neumann architecture. As the computational requirements for artificial intelligence grow 50 times faster than Moore’s Law for electronics, more unconventional approaches to computing and signal processing will appear on the horizon, and this roadmap will help identify future needs and challenges. In a very fertile field, experts in the field aim to present some of the dominant and most promising technologies for unconventional computing that will be around for some time to come. Within a holistic approach, the goal is to provide pathways for solidifying the field and guiding future impactful discoveries.

Finocchio, Giovanni (ORCID:0000000210433876)↗

Kokkos Array

The Kokkos Array library implements shared-memory array data structures and parallel task dispatch interfaces for data-parallel computational kernels that are performance-portable to multicore-CPU and manycore-accelerator (e.g., GPGPU) devices.

Edwards Daniel Sunderland, HaroldCarter↗

Digital video image system

Interactive recording and display device acts as very-high-speed data-input/output interface between analog (video) signals and standard digital-computer components. System can be used with various picture and memory sizes and can be controlled manually or by computer.

Neely, P. L.↗

Sonoluminescence Bubble Measurements using Vision-Based Algorithms

Vision-based measurement methods were used to measure bubble sizes in this sonoluminescence experiment. Bubble imaging was accomplished by placing the bubble between a bright light source and a microscope-CCD camera system. A collimated light-emitting diode was operated in a pulsed model with an adjustable time delay with respect to the piezo-electric transducer drive signal. The light-emitting diode produced a bubble shadowgraph consisting of a multiple exposure made by numerous light pulses imaged onto a charge-couple device camera. Each image was transferred from the camera to a computer-controlled machine vision system via a frame grabber. The frame grabber was equipped with on-board memory to accomodate sequential image buffering while images were transferred to the host processor and analyzed. This configuration allowed the host computer to perform diameter measurements, centroid position measurements and shape estimation in "real-time" as the next image was being acquired. Bubble size measurement accuracy with an uncertainty of 3 microns was achieved using standard lenses and machine vision algorithms. Bubble centroid position accuracy was also within the 3 micron tolerance of the vision system. This uncertainty estimation accounted for the optical spatial resolution, digitization errors and the edge detection algorithm accuracy. The vision algorithms include camera calibration, thresholding, edge detection, edge position determination, distance between two edges computations and centroid position computations.

Hall, Nancy R.↗

Active Flash: Out-of-core Data Analytics on Flash Storage

Next generation science will increasingly come to rely on the ability to perform efficient, on-the-fly analytics of data generated by high-performance computing (HPC) simulations, modeling complex physical phenomena. Scientific computing workflows are stymied by the traditional chaining of simulation and data analysis, creating multiple rounds of redundant reads and writes to the storage system, which grows in cost with the ever-increasing gap between compute and storage speeds in HPC clusters. Recent HPC acquisitions have introduced compute node-local flash storage as a means to alleviate this I/O bottleneck. We propose a novel approach, Active Flash, to expedite data analysis pipelines by migrating to the location of the data, the flash device itself. We argue that Active Flash has the potential to enable true out-of-core data analytics by freeing up both the compute core and the associated main memory. By performing analysis locally, dependence on limited bandwidth to a central storage system is reduced, while allowing this analysis to proceed in parallel with the main application. In addition, offloading work from the host to the more power-efficient controller reduces peak system power usage, which is already in the megawatt range and poses a major barrier to HPC system scalability. We propose an architecture for Active Flash, explore energy and performance trade-offs in moving computation from host to storage, demonstrate the ability of appropriate embedded controllers to perform data analysis and reduction tasks at speeds sufficient for this application, and present a simulation study of Active Flash scheduling policies. These results show the viability of the Active Flash model, and its capability to potentially have a transformative impact on scientific data analysis.

99 GENERAL AND MISCELLANEOUS↗

Kokkos Array

The Kokkos Array library implements shared-memory array data structures and parallel task dispatch interfaces for data-parallel computational kernels that are performance-portable to multicore-CPU and manycore-accelerator (e.g., GPGPU) devices.

Edwards Daniel Sunderland, Harold Carter↗

Thermodynamic origin of nonvolatility in resistive memory

Electronic switches based on the migration of high-density point defects, or memristors, are poised to revolutionize post-digital electronics. Despite significant research, key mechanisms for filament formation and oxygen transport remain unresolved, hindering our ability to predict and design device properties. For example, experiments have achieved 10 orders of magnitude longer retention times than predicted by current models. Here, using electrical measurements, scanning probe microscopy, and first-principles calculations on tantalum oxide memristors, we reveal that the formation and stability of conductive filaments crucially depend on the thermodynamic stability of the amorphous oxygen-rich and oxygen-poor compounds, which undergo composition phase separation. Including the previously neglected effects of this amorphous phase separation reconciles unexplained discrepancies in retention and enables predictive design of key performance indicators such as retention stability. Furthermore, this result emphasizes non-ideal thermodynamic interactions as key design criteria in post-digital devices with defect densities substantially exceeding those of today’s covalent semiconductors.

36 MATERIALS SCIENCE↗

Simulation of cosmic-ray induced soft errors and latchup in integrated-circuit computer memories

Soft errors have been induced in solid-state static RAM's by iron nuclei from the Lawrence Berkeley Laboratory (LBL) Bevalac, in experiments designed to prove the ability of iron-group cosmic rays to generate such errors. Subsequently, various delidded device types were tested in beams of argon and krypton ions from the LBL 88-inch Cyclotron, at energies near 2 MeV/nucleon. The latter tests showed that some devices are essentially immune to bit error while others are quite susceptible. Good agreement was obtained with model predictions in cases where the latter exist. Latchup, whose cause is attributed to individual heavy ions, was also observed in some device types.

Kolasinski, W. A.↗

Memristive linear algebra

The advent of memristive devices offers a promising avenue for efficient and scalable analog computing, particularly for linear algebra operations essential in various scientific and engineering applications. This paper investigates the potential of memristive crossbars in implementing matrix inversion algorithms. We explore both static and dynamic approaches, emphasizing the advantages of analog and in-memory computing for matrix operations beyond multiplication. In particular, we demonstrate that the electrical properties of memristive crossbars uniquely suit them for the evolution of a family of matrix exponentials, which can be exploited for the efficient computation of matrix inverses and online solutions for linear problems. Our results demonstrate that memristive arrays can reduce computational complexity. We also study power consumption and show a tradeoff between precision and energy. Furthermore, we address the challenges of device variability, precision, and scalability, providing insights into the practical implementation of these algorithms.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Engineering of Grain Boundaries in CeO 2 Enabling Tailorable Resistive Switching Properties

Abstract Defect engineering in valence change memories aimed at tuning the concentration and transport of oxygen vacancies are studied extensively, however mostly focusing on contribution from individual extended defects such as single dislocations and grain boundaries. In this work, the impact of engineering large numbers of grain boundaries on resistive switching mechanisms and performances is investigated. Three different grain morphologies, that is, “random network,” “columnar scaffold,” and “island‐like,” are realized in CeO 2 thin films. The devices with the three grain morphologies demonstrate vastly different resistive switching behaviors. The best overall resistive switching performance is shown in the devices with “columnar scaffold” morphology, where the vertical grain boundaries extending through the film facilitate the generation of oxygen vacancies as well as their migration under external bias. The observation of both interfacial and filamentary switching modes only in the devices with a “columnar scaffold” morphology further confirms the contribution from grain boundaries. In contrast, the “random network” or “island‐like” structures result in excessive or insufficient oxygen vacancy concentration migration paths. The research provides design guidelines for grain boundary engineering of oxide‐based resistive switching materials to tune the resistive switching performances for memory and neuromorphic computing applications.

36 MATERIALS SCIENCE↗

Analysis of self-heating of thermally assisted spin-transfer torque magnetic random access memory

Thermal assistance has been shown to significantly reduce the required operation power for spin torque transfer magnetic random access memory (STT-MRAM). Proposed heating methods include modified material stack compositions that result in increased self-heating or external heat sources. Here, we analyze the self-heating process of a standard perpendicular magnetic anisotropy STT-MRAM device through numerical simulations in order to understand the relative contributions of Joule, thermoelectric Peltier and Thomson, and tunneling junction heating. A 2D rotationally symmetric numerical model is used to solve the coupled electro-thermal equations including thermoelectric effects and heat absorbed or released at the tunneling junction. We compare self-heating for different common passivation materials, positive and negative electrical current polarity, and different device thermal anchoring and boundaries resistance configurations. The variations considered are found to result in significant differences in maximum temperatures reached. Average increases of 3 K, 10 K, and 100 K for different passivation materials, positive and negative polarity, and different thermal anchoring configurations, respectively, are observed. Furthermore, the highest temperatures, up to 424 K, are obtained for silicon dioxide as the passivation material, positive polarity, and low thermal anchoring with thermal boundary resistance configurations. Interestingly it is also found that due to the tunneling heat, Peltier effect, device geometry, and numerous interfacial layers around the magnetic tunnel junction (MTJ), most of the heat is dissipated on the lower potential side of the magnetic junction. We have observed this asymmetry in heating and is important as thermally assisted switching requires heating of the free layer specifically and this will be significantly different for the two polarity operations, set and reset.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

Autonomous Slat-Cove-Filler Device for Reduction of Aeroacoustic Noise Associated with Aircraft Systems

A slat cove filler is utilized to reduce airframe noise resulting from deployment of a leading edge slat of an aircraft wing. The slat cove filler is preferably made of a super elastic shape memory alloy, and the slat cove filler shifts between stowed and deployed shapes as the slat is deployed. The slat cove filler may be configured such that a separate powered actuator is not required to change the shape of the slat cove filler from its deployed shape to its stowed shape and vice-versa. The outer contour of the slat cove filler preferably follows a profile designed to maintain accelerating flow in the gap between the slat cove filler and wing leading edge to provide for noise reduction.

Turner, Travis L.↗