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 973 records · Page 54

FitCache: A Transparent Drop-In Framework for Multi-Tier Caching to Accelerate Distributed Deep Learning Workloads

Training in Deep learning (DL) remains highly compute- and data-intensive, with I/O becoming a critical bottleneck as models and datasets scale. Recent studies report that data loading can dominate training time, especially on large-scale HPC systems with shared parallel file systems (PFS). Existing caching approaches either rely on single-tier designs or require intrusive modifications to training pipelines, limiting their portability and effectiveness. In this work, we present FitCache, a transparent drop-in framework for multi-tier caching to accelerate distributed DL training by coordinating fast local memory (e.g., DRAM, Persistent Memory (PMem)) and NVMe as hierarchical caches atop PFS. Our design adapts to hardware diversity, i.e., if NVMe is missing, memory transparently acts as a caching tier, ensuring stable performance. FitCache transparently intercepts I/O requests and issues concurrent fetches across all tiers, returning data from the fastest responder without centralized metadata or static redirection paths. FitCache adapts to dynamic workloads and heterogeneous clusters while maintaining POSIX compatibility. Experiments on Frontier (2048 GPUs) and smaller research clusters show that FitCache reduces training time by up to 40% and per-batch I/O latency by up to 71.6% compared to Lustre Orion PFS, offering a drop-in solution for scalable DL training.

Hu, Guangxing [ORNL] (ORCID:0009000283203614)↗

Upsets in Erased Floating Gate Cells With High-Energy Protons

We discuss upsets in erased floating gate cells, due to large threshold voltage shifts, using statistical distributions collected on a large number of memory cells. The spread in the neutral threshold voltage appears to be too low to quantitatively explain the experimental observations in terms of simple charge loss, at least in SLC devices. The possibility that memories exposed to high energy protons and heavy ions exhibit negative charge transfer between programmed and erased cells is investigated, although the analysis does not provide conclusive support to this hypothesis.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

SUMC-DV hardware manual

The assembly, the physical and electrical characteristics, and the basic electrical tests of the Space Ultrareliable Modular Computer Demonstration Vehicle (SUMC-DV), are described. The descriptions include: (1) the packaging concepts, physical assembly, design and fabrication using design automation techniques of 10 different types of custom CMOS LSI arrays; (2) the fabrication and testing of the various components including the LSI arrays; (3) the hierarchy of the memory complement and the clock generation and distribution system; (4) system testing techniques; and (5) the procedure employed in the electrical checkout of the system.

Feller, A.↗

Parallel Conjugate Gradient: Effects of Ordering Strategies, Programming Paradigms, and Architectural Platforms

The Conjugate Gradient (CG) algorithm is perhaps the best-known iterative technique to solve sparse linear systems that are symmetric and positive definite. A sparse matrix-vector multiply (SPMV) usually accounts for most of the floating-point operations within a CG iteration. In this paper, we investigate the effects of various ordering and partitioning strategies on the performance of parallel CG and SPMV using different programming paradigms and architectures. Results show that for this class of applications, ordering significantly improves overall performance, that cache reuse may be more important than reducing communication, and that it is possible to achieve message passing performance using shared memory constructs through careful data ordering and distribution. However, a multi-threaded implementation of CG on the Tera MTA does not require special ordering or partitioning to obtain high efficiency and scalability.

Oliker, Leonid↗

Performance and usability enhancements for continuous subgraph matching queries on graph-structured data

A query graph, which includes vertices and edges, represents a query on graph-structured data. The query graph is decomposed into query subgraphs. A network analysis tool performs continuous subgraph matching queries to facilitate analysis of computer network traffic, social media events, or other streams of data represented as a dynamic data graph (graph-structured data). This can help identify emerging trends in the data. Some features of the network analysis tool enhance performance by effectively utilizing distributed computing resources (including processing cores and memory at different nodes of a cluster) to speed up the process of updating the dynamic data graph and detecting matches of query subgraphs. Features of a query graph building tool enhance usability by providing intuitive ways to specify query graphs and their subgraphs. Features of a results visualization tool enhance usability by providing an intuitive way to present the results of continuous subgraph matching queries.

Choudhury, Sutanay↗

Multiple transient memories in sheared suspensions: Robustness, structure, and routes to plasticity

Multiple transient memories, first discovered in charge-density-wave conductors, are a remarkable and initially counterintuitive example of how a system can store information about its driving. In this class of memories, a system can learn multiple driving inputs, nearly all of which are eventually forgotten despite their continual input. If sufficient noise is present, the system regains plasticity so that it can continue to learn new memories indefinitely. Recently, Keim and Nagel [Phys. Rev. Lett. 107, 010603 (2011)] showed how multiple transient memories could be generalized to a generic driven disordered system with noise, giving as an example simulations of a simple model of a sheared non-Brownian suspension. Here, we further explore simulation models of suspensions under cyclic shear, focusing on three main themes: robustness, structure, and overdriving. We show that multiple transient memories are a robust feature independent of many details of the model. The steady-state spatial distribution of the particles is sensitive to the driving algorithm; nonetheless, the memory formation is independent of such a change in particle correlations. Finally, we demonstrate that overdriving provides another means for controlling memory formation and retention.

42 ENGINEERING↗

Strongly coupling a cavity to inhomogeneous ensembles of emitters: Potential for long-lived solid-state quantum memories

We investigate theoretically the coupling of a cavity mode to a continuous distribution of emitters. We discuss the influence of the emitters' inhomogeneous broadening on the existence and on the coherence properties of the polaritonic peaks. We find that their coherence depends crucially on the shape of the distribution and not only on its width. Under certain conditions the coupling to the cavity protects the polaritonic states from inhomogeneous broadening, resulting in a longer storage time for a quantum memory based on emitter ensembles. When two different ensembles of emitters are coupled to the resonator, they support a peculiar collective dark state, which is also very attractive for the storage of quantum information.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

The Distribution of Lift Over Wing Tips and Ailerons

This investigation was carried out in the 5-foot wind tunnel of the Langley Memorial Aeronautical Laboratory for the purpose of obtaining more complete information on the distribution of lift between the ends of wing spars, the stresses in ailerons, and the general subject of airflow near the tip of a wing. It includes one series of tests on four models without ailerons, having square, elliptical, and raked tips respectively, and a second series of positively and negatively raked wings with ailerons adjusted to different settings. The results show that negatively raked tips give a more uniform distribution of air pressure than any of the other three arrangements, because the tip vortex does not disturb the flow at the trailing edge. Aileron loads are found to be less severe on wings with negative application to the calculation of aileron and wing stresses and also to facilitate the proper distribution of load in sand testing. Contour charts show in great detail the complex distribution lift over the wing.

Bacon, David L↗

Model reduction for a power grid model

We examine the complexity of constructing reduced order models for subsets of the variables needed to represent the state of the power grid. In particular, we apply model reduction techniques to the DeMarco-Zheng power grid model. We show that due to the oscillating nature of the solutions and the absence of timescale separation between resolved and unresolved variables, the construction of accurate reduced models becomes highly non-trivial because one has to account for long memory effects. In addition, we show that a reduced model that includes even a short memory is drastically better than a memoryless model.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Tuning the Mechanical Properties of Shape Memory Metallic Glass Composites with Brick and Mortar Designs

We use molecular dynamics simulations to characterize the tensile deformation and failure of shape memory alloy-bulk metallic glass composites (BMGCs). As the failure in BMGCs shifts from the MG matrix to the second phase different strategies are required to further improve their mechanical properties. Results show that a staggered brick BMGC displays enhanced ductility with mild strength loss following a homogeneously distributed plasticity. This demonstrates that judiciously designed BMGCs can effectively synergize the mechanical properties of metallic glasses and shape memory alloys by effectively distributing stress and preventing the generation of critical shear bands and premature failure while preserving strength.

36 MATERIALS SCIENCE↗

Time correlation functions in classical dense fluids

The velocity autocorrelation functions and memory functions of dense classical fluids may be directly obtained from the static radial distribution function g(r) in an approximate way. Following the Mori projection operator formalism, the memory functions may be related to the fluctuating force correlation. At low densities, these functions may be evaluated by following the trajectories of particle pairs in the interatomic potential. At higher densities, the force correlation functions can be evaluated approximately from particle pair trajectories via the potential of the mean force. The results for argon fluids at various densities and temperatures agree satisfactorily with the molecular dynamics and the Enskog values. The decrease of the diffusion coefficient with density is partly due to the nature of g(r) which reflects the stronger clustering of atoms at higher densities.

Tsang, T.↗

Quartile and Outlier Detection on Heterogeneous Clusters Using Distributed Radix Sort.

In the past few years, performance improvements in CPUs and memory technologies have outpaced those of storage systems. When extrapolated to the exascale, this trend places strict limits on the amount of data that can be written to disk for full analysis, resulting in an increased reliance on characterizing in-memory data. Many of these characterizations are simple, but require sorted data. This paper explores an example of this type of characterization - the identification of quartiles and statistical outliers - and presents a performance analysis of a distributed heterogeneous radix sort as well as an assessment of current architectural bottlenecks.

Spafford, Kyle L↗

Userspace Squash Filesystem for Launching Linux Containers on HPC Systems [Thesis]

The demand for user defined software stacks (UDSS) has been increasing in the high-performance computing (HPC) community. Container technology has become popular due to the flexibility and isolation it provides to HPC users. Container images must be available to all nodes involved for use in HPC and can be distributed to compute nodes in a variety of ways. A common method for container image distribution is to simply copy the container image to memory on each compute node, which can be time consuming at scale, and uses valuable memory on each node. The kernel mounted squash filesystem (squashfs) has proven fast and efficient for this task but requires root-level access. This paper will show a user space mounted squashfs is an efficient and secure solution for container image distribution.

97 MATHEMATICS AND COMPUTING↗

Accelerating shared file checkpoint with local burst buffers

A data management system and method for accelerating shared file checkpointing. Written application data is aggregated in an application data file created in a local burst buffer memory at a compute node, and an associated data mapping built index to maintain information related to the offsets into a shared file at which segments of the application data is to be stored in a parallel file system, and where in the buffer those segments are located. The node asynchronously transfers a data file containing the application data and the associated data mapping index to a file server for shared file storage. The data management system and method further accelerates shared file checkpointing in which a shared file, together with a map file that specifies how the shared file is to be distributed, is asynchronously transferred to local burst buffer memories at the nodes to accelerate reading of the shared file.

Gooding, Thomas↗

The effects of dose calculation resolution on dose accuracy for radiation therapy treatments of the lung. Part II. A comparison of dose distributions from an explicit lung model to dose distributions derived from a CT representation

Purpose: Due to limitations in computer memory and computation time, typical radiation therapy treatments are calculated with a voxel dimension on the order of several millimeters. The anatomy below this practical resolution is approximated as a homogeneous region uniform in atomic composition and density. The purpose of this article is to examine whether the exclusion of anatomic structure below the practical dose calculation resolution produces deviations in the resulting dose distributions. Methods: EGSnrc calculated dose distributions from the BRANCH lung model of Part I are compared and contrasted to dose distributions from a CT representation of the same BRANCH model for three different phases of the respiration cycle. Results: The exclusion of branching structures below a CT resolution of 1x1x2 mm{sup 3} resulted in a deviation in dose. The deviation in dose was as high as 14% but was localized around the branching structures. There was no significant variation in the dose deviation as a function of either field size or lung density. Conclusions: The exclusion of explicit branching structures of the lung in a CT representation creates localized deviations in dose. To ensure accurate dose calculations, CT resolution must be increased.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Quantum cryptography: individual eavesdropping with the knowledge of the error-correcting protocol

The quantum key distribution protocol BB84 combined with the repetition protocol for error correction is analysed from the point of view of its security against individual eavesdropping relying on quantum memory. It is shown that the mere knowledge of the error-correcting protocol changes the optimal attack and provides the eavesdropper with additional information on the distributed key. (fifth seminar in memory of d.n. klyshko)

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Comprehensive Study of the Polarization Behavior of LiFePO 4 Electrodes Based on a Many-Particle Model

The unique polarization behavior of LiFePO 4 electrodes was investigated using mathematical simulations with a many-particle model, which have a non-monotonic potential profile for each LiFePO 4 particle, and analyzed by the active population concept. The present simulations reveal two known polarization behaviors, namely the memory effect and path dependence. Notably, a hitherto unknown polarization behavior, the so-called relaxation-induced polarization (RIP), was also identified. The memory effect requires, at a minimum, a sequential four-step operation. By comparison, RIP is triggered only by a one-step operation, namely a long rest. In effect, the polarization associated with the memory effect and RIP was caused by a reduction of the active particles in the two-phase region via relaxation during a rest. The path-dependence mechanism was attributed to kinetically inhomogeneous reactions for each particle. We further found that narrowing the particle size distribution was an effective means to reduce polarization viz. the memory effect and path dependence of LiFePO 4 electrodes. However, RIP could not be suppressed by narrowing the particle size distribution. Furthermore, a comprehensive understanding of these polarization behaviors with our model provides a more accurate way to estimate the state of charge in Li-ion batteries with LiFePO 4 electrodes.

25 ENERGY STORAGE↗

SU-F-BRF-01: A GPU Framework for Developing Interactive High-Resolution Patient-Specific Biomechanical Models

Purpose: To develop a GPU-based framework that can generate highresolution and patient-specific biomechanical models from a given simulation CT and contoured structures, optimized to run at interactive speeds, for addressing adaptive radiotherapy objectives. Method: A Massspring-damping (MSD) model was generated from a given simulation CT. The model's mass elements were generated for every voxel of anatomy, and positioned in a deformation space in the GPU memory. MSD connections were established between neighboring mass elements in a dense distribution. Contoured internal structures allowed control over elastic material properties of different tissues. Once the model was initialized in GPU memory, skeletal anatomy was actuated using rigid-body transformations, while soft tissues were governed by elastic corrective forces and constraints, which included tensile forces, shear forces, and spring damping forces. The model was validated by applying a known load to a soft tissue block and comparing the observed deformation to ground truth calculations from established elastic mechanics. Results: Our analyses showed that both local and global load experiments yielded results with a correlation coefficient R{sup 2} > 0.98 compared to ground truth. Models were generated for several anatomical regions. Head and neck models accurately simulated posture changes by rotating the skeletal anatomy in three dimensions. Pelvic models were developed for realistic deformations for changes in bladder volume. Thoracic models demonstrated breast deformation due to gravity when changing treatment position from supine to prone. The GPU framework performed at greater than 30 iterations per second for over 1 million mass elements with up to 26 MSD connections each. Conclusions: Realistic simulations of site-specific, complex posture and physiological changes were simulated at interactive speeds using patient data. Incorporating such a model with live patient tracking would facilitate real time assessment of variations of the actual anatomy and delivered dose for adaptive intervention and re-planning.

60 APPLIED LIFE SCIENCES↗