Search NASA⌕ Search

SEARCH · Search NASA

Results for “distributed 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 1,009 records · Page 56

Differential Power Processing for Ultra-Efficient Data Storage

Here this paper presents the hardware, software, and power codesign of an ultra-efficient data storage server with differential power processing (DPP). DPP can reduce the power conversion stress, improve the efficiency, and enhance the functionality of modular power electronics systems. The power inputs of a large number of hard disk drives (HDDs) were connected in series and supported by a multiport ac-coupled differential power processing (MAC-DPP) converter through a multiwinding transformer. Methods for controlling the multi-input multi-output power flow in the multiwinding transformer while avoiding core saturation were investigated. A ten-port MAC-DPP prototype with 700-W/in 3 power density was built to support a 450-W HDD storage system with ten series-stacked voltage domains. The prototype was tested on a 50-HDD server testbench, and the overall system loss is below 1 W (99.77% system efficiency). The server was able to maintain high-speed reading and writing operation of all 50 HDDs against the worst hot-swapping scenarios. A variety of hardware/software configurations and many cloud storage techniques were tested on the fully functioning server. Experimental results show that the energy efficiency of large-scale information systems (CPU/GPU clusters, memory banks, HDD arrays, etc.) can be greatly improved by software, hardware, and power codesign.

42 ENGINEERING↗

Linking Deep and Shallow Convective Mass Fluxes via an Assumed Entrainment Distribution in CAM5‐CLUBB: Parameterization and Simulated Precipitation Variability

Abstract We modify the Zhang‐McFarlane (ZM) deep convection scheme in the Community Atmosphere Model version 5 to couple it with a unified parameterization for boundary‐layer turbulence and shallow convection, that is, Cloud Layers Unified by Binormals (CLUBB). By assuming a lognormal distribution of entrainment rate across the entire moist convective regimes, we link mass fluxes between shallow and deep convection, which are partitioned by the entrainment rate of the shallowest deep convective plume. Hence, a new deep convective closure is established which is coupled to the sub‐grid vertical motion variability in CLUBB. The convection feedback (or memory) effects are also considered to decrease the entrainment spectrum width and enhance the vertical velocity variability that further affect deep convection. Results show that the revised scheme improves the precipitation simulations in terms of the mean state and variability at various timescales, such as the alleviated double‐intertropical convergence zone and more realistic simulations of the seasonal variation of monsoon precipitation over East Asia, Madden‐Julian Oscillation, and precipitation diurnal phase propagations downstream of large terrains. The improvements are still seen in many aspects such as the mean‐state precipitation when turning off the convection feedback impacts in the revised scheme, emphasizing the benefits of using the modified mass‐flux closure. However, the convection feedbacks have considerable effects on the precipitation diurnal cycle simulations over regions with late‐afternoon precipitation peaks. Overall, the revised scheme provides a unified treatment for sub‐grid vertical motions across regimes of boundary‐layer turbulence, shallow convection, and deep convection, leading to better‐simulated precipitation at various timescales.

54 ENVIRONMENTAL SCIENCES↗

System and method for implementing periodic early discard in on-chip buffer memories of network elements

An advance is made over the prior art in accordance with the principles of the present invention that is directed to a new approach for a system and method for a buffer management scheme called Periodic Early Discard (PED). The invention builds on the observation that, in presence of TCP traffic, the length of a queue can be stabilized by selection of an appropriate frequency for packet dropping. For any combination of number of TCP connections and distribution of the respective RTT values, there exists an ideal packet drop frequency that prevents the queue from over-flowing or under-flowing. While the value of the ideal packet drop frequency may quickly change over time and is sensitive to the series of TCP connections affected by past packet losses, and most of all is impossible to compute inline, it is possible to approximate it with a margin of error that allows keeping the queue occupancy within a pre-defined range for extended periods of time. The PED scheme aims at tracking the (unknown) ideal packet drop frequency, adjusting the approximated value based on the evolution of the queue occupancy, with corrections of the approximated packet drop frequency that occur at a timescale that is comparable to the aggregate time constant of the set of TCP connections that traverse the queue.

97 MATHEMATICS AND COMPUTING↗

Risso's dolphins plan foraging dives

ABSTRACT Humans remember the past and use that information to plan future actions. Lab experiments that test memory for the location of food show that animals have a similar capability to act in anticipation of future needs, but less work has been done on animals foraging in the wild. We hypothesized that planning abilities are critical and common in breath-hold divers who adjust each dive to forage on prey varying in quality, location and predictability within constraints of limited oxygen availability. We equipped Risso's dolphins with sound-and-motion recording tags to reveal where they focus their attention through their externally observable echolocation and how they fine tune search strategies in response to expected and observed prey distribution. The information from the dolphins was integrated with synoptic prey data obtained from echosounders on an underwater vehicle. At the start of the dives, whales adjusted their echolocation inspection ranges in ways that suggest planning to forage at a particular depth. Once entering a productive prey layer, dolphins reduced their search range comparable to the scale of patches within the layer, suggesting that they were using echolocation to select prey within the patch. On ascent, their search range increased, indicating that they decided to stop foraging within that layer and started searching for prey in shallower layers. Information about prey, learned throughout the dive, was used to plan foraging in the next dive. Our results demonstrate that planning for future dives is modulated by spatial memory derived from multi-modal prey sampling (echoic, visual and capture) during earlier dives.

Arranz, Patricia↗

THE FORMATION OF YOUNG DENSE STAR CLUSTERS THROUGH MERGERS

Young star clusters such as NGC 3603 and Westerlund 1 and 2 in the Milky Way and R136 in the Large Magellanic Cloud are dynamically more evolved than expected based on their current relaxation times. In particular, the combination of a high degree of mass segregation, a relatively low central density, and the large number of massive runaway stars in their vicinity are hard to explain with the monolithic formation of these clusters. Young star clusters can achieve such a mature dynamical state if they formed through the mergers of a number of less massive clusters. The shorter relaxation times of less massive clusters cause them to dynamically evolve further by the time they merge, and the merger product preserves the memory of the dynamical evolution of its constituent clusters. With a series of N-body simulations, we study the dynamical evolution of single massive clusters and those that are assembled through merging smaller clusters together. We find that the formation of massive star clusters through the mergers of smaller clusters can reproduce the currently observed spatial distribution of massive stars, the density, and the characteristics (number and mass distribution) of the stars ejected as runaways from young dense clusters. We therefore conclude that these clusters and possibly other young massive star clusters formed through the mergers of smaller clusters.

79 ASTRONOMY AND ASTROPHYSICS↗

A Subspace-Inclusive Sampling Method for the Computational Design of Compositionally Graded Alloys

Abstract Compositionally graded alloys, a subclass of functionally graded materials (FGMs), utilize localized variations in composition with a single metal part to achieve higher performance than traditional single material parts. In previous work [Kirk, T., Galvan, E., Malak, R., and Arroyave, R., 2018, “Computational Design of Gradient Paths in Additively Manufactured Functionally Graded Materials,” J. Mech. Des., 140, p. 111410. 10.1115/1.4040816], the authors presented a computational design methodology that avoids common issues which limit a gradient alloy’s feasibility, such as deleterious phases, and optimizes for performance objectives. However, the previous methodology only samples the interior of a composition space, meaning designed gradients must include all elements in the space throughout the gradient. Because even small amounts of additional alloying elements can introduce new deleterious phases, this characteristic often neglects potentially simpler solutions to otherwise unsolvable problems and, consequently, discourages the addition of new elements to the state space. The present work improves upon the previous methodology by introducing a sampling method that includes subspaces with fewer elements in the design search. The new method samples within an artificially expanded form of the state space and projects samples outside the true region to the nearest true subspace. This method is evaluated first by observing the sample distribution in each subspace of a 3D, 4D, and 5D state space. Next, a parametric study in a synthetic 3D problem compares the performance of the new sampling scheme to the previous methodology. Lastly, the updated methodology is applied to design a gradient from stainless steel to equiatomic NiTi that has practical uses such as embedded shape memory actuation and for which the previous methodology fails to find a feasible path.

Engineering↗

Asymptotic Behavior of Memristive Circuits

The interest in memristors has risen due to their possible application both as memory units and as computational devices in combination with CMOS. This is in part due to their nonlinear dynamics, and a strong dependence on the circuit topology. We provide evidence that also purely memristive circuits can be employed for computational purposes. In the present paper we show that a polynomial Lyapunov function in the memory parameters exists for the case of DC controlled memristors. Such a Lyapunov function can be asymptotically approximated with binary variables, and mapped to quadratic combinatorial optimization problems. This also shows a direct parallel between memristive circuits and the Hopfield-Little model. In the case of Erdos-Renyi random circuits, we show numerically that the distribution of the matrix elements of the projectors can be roughly approximated with a Gaussian distribution, and that it scales with the inverse square root of the number of elements. This provides an approximated but direct connection with the physics of disordered system and, in particular, of mean field spin glasses. Using this and the fact that the interaction is controlled by a projector operator on the loop space of the circuit. We estimate the number of stationary points of the approximate Lyapunov function and provide a scaling formula as an upper bound in terms of the circuit topology only.

97 MATHEMATICS AND COMPUTING↗

Actuation method and apparatus, micropump, and PCR enhancement method

An actuation apparatus includes at least one magnetic shape memory (MSM) element containing a material configured to expand and/or contract in response to exposure to a magnetic field. Among other things, the MSM element may be configured to pump fluid through a micropump by expanding and/or contracting in response to the magnetic field. The magnetic field may rotate about an axis of rotation and exhibit a distribution having a component substantially perpendicular to the axis of rotation. Further, the magnetic field distribution may include at least two components substantially orthogonal to one another lying in one or more planes perpendicular to the axis of rotation. The at least one MSM element may contain nickel, manganese, and gallium. A polymerase chain reaction (PCR) may be enhanced by contacting a PCR reagent and DNA material with the MSM element.

36 MATERIALS SCIENCE↗

Improved Evaluation of Large Network Matrices for Linear Power Flow Within Optimization Problems: Preprint

This work discusses methods for evaluating the Power Transfer Distribution Factor (PTDF) and Line Outage Distribution Factor (LODF) matrices by employing sparse linear algebra for large-scale computing applications. These matrices are critical in many power systems applications, such as the Unit Commitment Problem (UC), pre- and post-contingency power flow analysis, and transmission expansion. These matrices are typically dense, which means they require a significant amount of time and memory to be computed for large networks. However, by analyzing the structure of the matrices and their computation method, it is possible to use reduced memory methods based on sparse matrix operations. This paper shows that sparse linear algebra algorithms are faster and require less memory and time than traditional dense approaches. Additionally, we explore the effect of matrix sparsification by eliminating trailing digits on power flow calculations.

ENERGY PLANNING, POLICY, AND ECONOMY↗

Formation of the Conducting Filament in TaO x -Resistive Switching Devices by Thermal-Gradient-Induced Cation Accumulation

Here, the distribution of tantalum and oxygen ions in electroformed and/or switched TaO x -based resistive switching devices has been assessed by high-angle annular dark-field microscopy, X-ray energy-dispersive spectroscopy, and electron energy-loss spectroscopy. The experiments have been performed in the plan-view geometry on the cross-bar devices producing elemental distribution maps in the direction perpendicular to the electric field. The maps revealed an accumulation of +20% Ta in the inner part of the filament with a 3.5% Ta-depleted ring around it. The diameter of the entire structure was approximately 100 nm. The distribution of oxygen was uniform with changes, if any, below the detection limit of 5%. We interpret the elemental segregation as due to diffusion driven by the temperature gradient, which in turn is induced by the spontaneous current constriction associated with the negative differential resistance-type I–V characteristics of the as-fabricated metal/oxide/metal structures. A finite-element model was used to evaluate the distribution of temperature in the devices and correlated with the elemental maps. In addition, a fine-scale (~5 nm) intensity contrast was observed within the filament and interpreted as due phase separation of the functional oxide in the two-phase composition region. Understanding the temperature-gradient-induced phenomena is central to the engineering of oxide memory cells.

36 MATERIALS SCIENCE↗

$\mathrm{SageNet}$: Fast Neural Network Emulation of the Stiff-amplified Gravitational Waves from Inflation

Accurate modeling of the inflationary gravitational waves (GWs) requires time-consuming, iterative numerical integrations of differential equations to take into account their backreaction on the expansion history. To improve computational efficiency while preserving accuracy, we present the Stiff-amplified Gravitational-wave Emulator Network (SageNet), a deep learning framework designed to replace conventional numerical solvers (code available at https://github.com/YifangLuo/SageNet). SageNet employs a long short-term memory architecture to emulate the present-day energy density spectrum of the inflationary GWs with possible stiff amplification, Ω GW (f). Trained on a data set of 25,689 numerically generated solutions, SageNet allows accurate reconstructions of Ω GW (f) and generalizes well to a wide range of cosmological parameters; 90.9% of the test emulations with randomly distributed parameters exhibit errors of under 4%. In addition, SageNet demonstrates its ability to learn and reproduce the artificial, adaptive sampling patterns in numerical calculations, which implement denser sampling of frequencies around changes in spectral indices in Ω GW (f). The dual capability of learning both physical and artificial features of the numerical GW spectra establishes SageNet as a robust alternative to exact numerical methods. Finally, our benchmark tests show that SageNet reduces the computation time from tens of seconds to milliseconds, achieving a speedup of ∼10 4 times over standard CPU-based numerical solvers with the potential for further acceleration on GPU hardware. These capabilities make SageNet a powerful tool for accelerating Bayesian inference procedures for extended cosmological models. In a broad sense, the SageNet framework offers a fast, accurate, and generalizable solution to modeling cosmological observables whose theoretical predictions demand costly differential equation solvers.

Astronomy data modeling↗

Design of a neutral thermal scattering (NeTS) module for hydrogen in light water

The accurate representation of thermal scattering law (TSL) data is integral to the design and characterization of many modern nuclear systems, particularly those using light water as a moderator/coolant. As a material-dependent distribution over energy-momentum phase space, the TSL may exhibit a variety of unique and relevant conditional (e.g., temperature, pressure) and compositional (e.g., porosity, stoichiometry, radiation damage) dependencies. Currently, there are various approaches to incorporate temperature dependence, which is especially important in coupled neutronic-thermal hydraulic simulations. In each approach, there is an inherent tradeoff between memory consumption and accuracy. Some techniques require tens to hundreds of MBs or more, while others fail to reproduce the underlying data to within 10% error over the considered input domain, despite having a reduced storage burden. This work aims to address both sides of the tradeoff simultaneously by implementing a novel deep learning (DL) approach to TSL representation. The neural thermal scattering (NeTS) concept, which is amenable to an arbitrary number of dependencies, is demonstrated via the inclusion of temperature dependence into a highly compact, highly accurate functional form of the multi- variate TSL (i.e., S(α, β, T)) for hydrogen in light water. Resulting storage requirements are on the order of 100 kB, and median and maximum percent deviations are on the order of 0.1% and 1%, respectively. These measures represent a step improvement over previous techniques. Notably, the developed neural network and feature methodology build on those employed in prior work on beryllium oxide. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Progressive Tree-Based Compression of Large-Scale Particle Data

Scientific simulations and observations using particles have been creating large datasets that require effective and efficient data reduction to store, transfer, and analyze. However, current approaches either compress only small data well while being inefficient for large data, or handle large data but with insufficient compression. Toward effective and scalable compression/decompression of particle positions, we introduce new kinds of particle hierarchies and corresponding traversal orders that quickly reduce reconstruction error while being fast and low in memory footprint. Our solution to compression of large-scale particle data is a flexible block-based hierarchy that supports progressive, random-access, and error-driven decoding, where error estimation heuristics can be supplied by the user. For low-level node encoding, we introduce new schemes that effectively compress both uniform and densely structured particle distributions. Our proposed methods thus target all three phases of a tree-based particle compression pipeline, namely tree construction, tree traversal, and node encoding. In conclusion, the improved efficacy and flexibility of these methods over existing compressors are demonstrated through extensive experimentation, using a wide range of scientific particle datasets.

97 MATHEMATICS AND COMPUTING↗

Non-Markovian relaxation spectroscopy of fluxonium qubits

Recent studies have shown that parasitic two-level systems (TLS) in superconducting qubits, which are a leading source of decoherence, can have relaxation times longer than the qubits themselves. However, the standard techniques used to characterize qubit relaxation is only valid for measuring T 1 under the Born-Markov approximation and could mask environmental memory effects in practice. Here, we introduce two-timescale relaxometry, a technique to probe the qubit and environment relaxation simultaneously and efficiently. We apply it to high-coherence fluxonium qubits over a frequency range of 0.1-0.4 GHz, and reveal a discrete spectrum of TLS with millisecond lifetimes. Our analysis of the spectrum is consistent with a random distribution of TLS in the aluminum oxide tunnel barrier of the Josephson junction chain of the fluxonium, with a spectral and volumetric density and average electric dipole similar to previous TLS studies at much higher frequencies. Our study suggests that investigating and mitigating TLS in the junction chain is crucial to the development of various types of noise-protected qubits in circuit QED.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Model fusion with physics-guided machine learning: Projection-based reduced-order modeling

The unprecedented amount of data generated from experiments, field observations, and large-scale numerical simulations at a wide range of spatiotemporal scales has enabled the rapid advancement of data-driven and especially deep learning models in the field of fluid mechanics. Although these methods are proven successful for many applications, there is a grand challenge of improving their generalizability. This is particularly essential when data-driven models are employed within outer-loop applications like optimization. In this work, we put forth a physics-guided machine learning (PGML) framework that leverages the interpretable physics-based model with a deep learning model. Leveraging a concatenated neural network design from multi-modal data sources, the PGML framework is capable of enhancing the generalizability of data-driven models and effectively protects against or inform about the inaccurate predictions resulting from extrapolation. We apply the PGML framework as a novel model fusion approach combining the physics-based Galerkin projection model and long- to short-term memory (LSTM) network for parametric model order reduction of fluid flows. We demonstrate the improved generalizability of the PGML framework against a purely data-driven approach through the injection of physics features into intermediate LSTM layers. Our quantitative analysis shows that the overall model uncertainty can be reduced through the PGML approach, especially for test data coming from a distribution different than the training data. Moreover, we demonstrate that our approach can be used as an inverse diagnostic tool providing a confidence score associated with models and observations. The proposed framework also allows for multi-fidelity computing by making use of low-fidelity models in the online deployment of quantified data-driven models.

42 ENGINEERING↗

Design and Implementation of Papyrus: Parallel Aggregate Persistent Storage

A surprising development in recently announced HPC platforms is the addition of, sometimes massive amounts of, persistent (nonvolatile) memory (NVM) in order to increase memory capacity and compensate for plateauing I/O capabilities. However, there are no portable and scalable programming interfaces using aggregate NVM effectively. This paper introduces Papyrus: a new software system built to exploit emerging capability of NVM in HPC architectures. Papyrus (or Parallel Aggregate Persistent -YRU- Storage) is a novel programming system that provides features for scalable, aggregate, persistent memory in an extreme-scale system for typical HPC usage scenarios. Papyrus mainly consists of Papyrus Virtual File System (VFS) and Papyrus Template Container Library (TCL). Papyrus VFS provides a uniform aggregate NVM storage image across diverse NVM architectures. It enables Papyrus TCL to provide a portable and scalable high-level container programming interface whose data elements are distributed across multiple NVM nodes without requiring the user to handle complex communication, synchronization, replication, and consistency model. We evaluate Papyrus on two HPC systems, including UTK Beacon and NERSC Cori, using real NVM storage devices.

Kim, Jungwon↗

Design of functional materials based on new principles of disorder. Final Report

Disorder is present to some extent in every material. The disorder can be sparse and localized – individual defects – or globally such as in a glass. Normally disorder is considered to be detrimental to function. However, this is not necessarily always the case. This project addressed the question: How does disorder provide new potentialities for the behavior of a material that would be absent in a crystalline material of the same composition? We have discovered a new principle for disordered matter: Independence of bond-level response. This work has introduced the possibility of creating new classes of mechanical meta-materials with applications everywhere from microscopic biological molecules (allostery) to large scale architectural disordered networks. During the course of this grant, we have subsequently elaborated on these novel ideas about bond-level independence of properties in disordered materials. In particular, this study has required that we understand the limits of disorder. That is, if crystals are perfect order, is there another diametrically opposed pole which is the most disordered state? We have shown that, in a well-defined way, jamming is one extreme disordered pole of rigid matter. Jamming is a far-from-equilibrium way of creating a rigid solid by rapidly quenching a random distribution of particles to a mechanically stable system at zero temperature. To understand the jammed state of matter, we have elucidated the low-frequency excitation of these systems. This has allowed a new way of understanding materials that cannot be easily classified as crystals or glasses. These ideas can be extended to address the mechanical properties of polycrystalline matter. Memory formation in matter can take inspiration from memory that occurs in a biological context. The issue of memory formation raises another set of questions at the heart of how the energy landscape is organized and how it can be manipulated to store information in a solid. This project has explored how to understand the degree to which multiple memories can be stored in a single system and has created a holistic understanding of how inputs can be used to create a desired response.

36 MATERIALS SCIENCE↗

Optimizing Distributed Training on Frontier for Large Language Models

Large language models (LLMs) have demonstrated remarkable success as foundational models, benefiting various downstream applications through fine-tuning. Loss scaling studies have demonstrated the superior performance of larger LLMs compared to their smaller counterparts. Nevertheless, training LLMs with billions of parameters poses significant challenges and requires considerable computational resources. For example, training a one trillion parameter GPT-style model on 20 trillion tokens requires a staggering 120 million exaflops. This research explores efficient distributed training strategies to extract this computation from Frontier, the world's first exascale supercomputer. We enable and investigate various model and data parallel training techniques, such as tensor parallelism, pipeline parallelism, and sharded data parallelism, to facilitate training a trillion-parameter model on Frontier. We empirically assess these techniques and their associated parameters to determine their impact on memory footprint, communication latency, and GPU's computational efficiency. We analyze the complex interplay among these techniques and find a strategy to combine them to achieve high throughput through hyperparameter tuning. We have identified efficient strategies for training large LLMs of varying sizes through empirical analysis and hyperparameter tuning. For 22 Billion, 175 Billion, and 1 Trillion parameters, we achieved GPU throughputs of 38.38%, 36.14%, and 31.96%, respectively. For the training of the 175 Billion parameter model and the 1 Trillion parameter model, we achieved 100% weak scaling efficiency on 1024 and 3072 Mi250X GPUs, respectively. We also achieved strong scaling efficiencies of 89% and 87% for these two models. We trained these models only tens of iterations instead of training till completion.

Yin, Junqi↗