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 541 records · Page 30

rustpix

rustpix is a high-performance, open-source Rust library with first-class Python bindings (via PyO3) for processing pixel-detector data in neutron imaging. It targets time-stamping detectors such as Timepix3 (TPX3) at ORNL's Spallation Neutron Source (VENUS beamline), where each detected neutron deposits charge across a cluster of pixels within a very high-rate event stream (96M+ hits/sec). rustpix parses TPX3 event data in parallel using memory-mapped I/O, offers four interchangeable clustering algorithms (ABS adjacency-based search, DBSCAN, graph/union-find connected components, and a parallel grid method), and extracts weighted, super-resolved centroids to produce neutron-event lists. A streaming architecture lets it process files larger than available memory. rustpix is distributed as a pip-installable Python package (with NumPy integration), Rust crates, a command-line tool, and an interactive GUI; it writes HDF5, Apache Arrow, and CSV; and it is designed to extend to TPX4 and other detector types. Released as open-source under the MIT License.

Zhang, Chen [Oak Ridge National Laboratory (ORNL),↗

Secure VM for Monitoring Industrial Process Controllers

In this paper, we examine the biological immune system as an autonomic system for self-protection, which has evolved over millions of years probably through extensive redesigning, testing, tuning and optimization process. The powerful information processing capabilities of the immune system, such as feature extraction, pattern recognition, learning, memory, and its distributive nature provide rich metaphors for its artificial counterpart. Our study focuses on building an autonomic defense system, using some immunological metaphors for information gathering, analyzing, decision making and launching threat and attack responses. In order to detection Stuxnet like malware, we propose to include a secure VM (or dedicated host) to the SCADA Network to monitor behavior and all software updates. This on-going research effort is not to mimic the nature but to explore and learn valuable lessons useful for self-adaptive cyber defense systems.

99 GENERAL AND MISCELLANEOUS↗

An Adaptive-Mesh-Refinement Based Computational Tool for Simulating Catalysis at Mesoscale

In this work, we present a computational tool for mesoscale applications using open-source exascale- computing compatible adaptive-mesh-refinement (AMR) library, AMReX [2]. AMReX is software library that enables development of application solvers with block-structured Cartesian AMR. Our tool has capabilities to include realistic geometry representation, chemical species transport, reactions and thermodynamics that are critical for capturing mesoscale physics. A significant achievement is the ability of our solver to automatically import electron microscopy data in the form of a stereolithography (STL) or pixelated file format (mrc, tiff) without undergoing the tedious task of unstructured mesh generation. This feature allows for rapid simulation of catalyst particles with complex morphologies using an immersed-boundary formulation. The use of AMR allows for higher resolutions at catalyst surface interfaces, which in turn provides an accurate description of surface reactions and transport. Our solver uses a hybrid distributed and shared memory parallelism (OpenMP/GPU-based) with which strong scaling up to 10,000 processors for realistic catalyst particle simulations have been demonstrated.

BIOMASS FUELS,MATHEMATICS AND COMPUTING↗

Access of named data elements in coordination namespace

An approach is described that provides access to a named data element in a Coordination Namespace that is stored in a memory that is distributed amongst a set of nodes. A request of a name corresponding to the named data element is received from a requesting process and the approach responsively searches for the name in the Coordination Namespace. In response to determining an absence of data corresponding to the named data element, a pending state is indicated to the requesting process. In response to determining that the data corresponding to the named data element exists, a successful state is returned to the requesting process. In one embodiment, the successful state also includes providing the requesting process with access to the data corresponding to the named data element.

97 MATHEMATICS AND COMPUTING↗

Parallel 3-D method of characteristics in MPACT

A new parallel 3-D MOC kernel has been developed and implemented in MPACT which makes use of the modular ray tracing technique to reduce computational requirements and to facilitate parallel decomposition. The parallel model makes use of both distributed and shared memory parallelism which are implemented with the MPI and OpenMP standards, respectively. The kernel is capable of parallel decomposition of problems in space, angle, and by characteristic rays up to 0(104) processors. Initial verification of the parallel 3-D MOC kernel was performed using the Takeda 3-D transport benchmark problems. The eigenvalues computed by MPACT are within the statistical uncertainty of the benchmark reference and agree well with the averages of other participants. The MPACT k{sub eff} differs from the benchmark results for rodded and un-rodded cases by 11 and -40 pcm, respectively. The calculations were performed for various numbers of processors and parallel decompositions up to 15625 processors; all producing the same result at convergence. The parallel efficiency of the worst case was 60%, while very good efficiency (>95%) was observed for cases using 500 processors. The overall run time for the 500 processor case was 231 seconds and 19 seconds for the case with 15625 processors. Ongoing work is focused on developing theoretical performance models and the implementation of acceleration techniques to minimize the number of iterations to converge. (authors)

97 MATHEMATICS AND COMPUTING↗

Software-Hardware Co-design of Heterogeneous SmartNIC System for Recommendation Models Inference and Training

Deep Learning Recommendation Models (DLRMs) are critical applications in various domains and have evolved as one of the single largest machine learning applications. Trillions of DLRM parameters exceed the on-chip memory capacity of GPUs. Large-scale multi-node systems are required for distributed DLRM inference and training, which suffer from the all-to-all communication bottleneck, mainly limiting the scalability of ever-growing DLRMs. In recent years, SmartNICs have evolved with coupled computation and communication capabilities providing opportunities for a powerful heterogeneous device in the system. However, there isn't such a distributed system that fully leverages the abundant smartNIC resources that resolve the scalability issue of DLRMs. In this work, we proposed a software-hardware co-design of a heterogeneous smartNIC system that resolves the communication bottleneck of distributed DLRMs, mitigates the memory bandwidth pressure, and improves computation efficiency. We provide a set of smartNIC designs of cache systems (including local cache and remote cache) and smartNIC computation kernels which reduce data movement, relieve memory lookup intensity, and improve the GPU's computation efficiency. In addition, we propose a graph algorithm that improves the data locality of queries within batches which optimizes the overall system performance with higher data reuse. Our evaluation shows that our system achieves 2.1x latency speedup for inference and 1.6x throughput speedup for training.

Guo, Anqi↗

RDMA vs. RPC for Implementing Distributed Data Structures

Distributed data structures are key to implementing scalable applications for scientific simulations and data analysis. In this paper we look at two implementation styles for distributed data structures: remote direct memory access (RDMA) and remote procedure call (RPC). We focus on operations that require individual accesses to remote portions of a distributed data structure, e.g., accessing a hash table bucket or distributed queue, rather than global operations in which all processors collectively exchange information. We look at the trade-offs between the two styles through microbenchmarks and a performance model that approximates the cost of each. The RDMA operations have direct hardware support in the network and therefore lower latency and overhead, while the RPC operations are more expressive but higher cost and can suffer from lack of attentiveness from the remote side. We also run experiments to compare the real-world performance of RDMA- and RPC-based data structure operations with the predicted performance to evaluate the accuracy of our model, and show that while the model does not always precisely predict running time, it allows us to choose the best implementation in the examples shown. We believe this analysis will assist developers in designing data structures that will perform well on current network architectures, as well as network architects in providing better support for this class of distributed data structures.

distributed data structures↗

Hierarchical resilience with lightweight threads.

This paper proposes methodology for providing robustness and resilience for a highly threaded distributed- and shared-memory environment based on well-defined inputs and outputs to lightweight tasks. These inputs and outputs form a failure 'barrier', allowing tasks to be restarted or duplicated as necessary. These barriers must be expanded based on task behavior, such as communication between tasks, but do not prohibit any given behavior. One of the trends in high-performance computing codes seems to be a trend toward self-contained functions that mimic functional programming. Software designers are trending toward a model of software design where their core functions are specified in side-effect free or low-side-effect ways, wherein the inputs and outputs of the functions are well-defined. This provides the ability to copy the inputs to wherever they need to be - whether that's the other side of the PCI bus or the other side of the network - do work on that input using local memory, and then copy the outputs back (as needed). This design pattern is popular among new distributed threading environment designs. Such designs include the Barcelona STARS system, distributed OpenMP systems, the Habanero-C and Habanero-Java systems from Vivek Sarkar at Rice University, the HPX/ParalleX model from LSU, as well as our own Scalable Parallel Runtime effort (SPR) and the Trilinos stateless kernels. This design pattern is also shared by CUDA and several OpenMP extensions for GPU-type accelerators (e.g. the PGI OpenMP extensions).

43 PARTICLE ACCELERATORS↗

A computationally efficient ductile damage model accounting for nucleation and micro-inertia at high triaxialities

The computational formulation of a micro-mechanical material model for the dynamic failure of ductile metals is presented in this paper. The statistical nature of porosity initiation is accounted for by introducing an arbitrary probability density function which describes the pores nucleation pressures. Each micropore within the representative volume element is modeled as a thick spherical shell made of plastically incompressible material. The treatment of porosity by a distribution of thick-walled spheres also allows for the inclusion of micro-inertia effects under conditions of shock and dynamic loading. The second order ordinary differential equation governing the microscopic porosity evolution is solved with a robust implicit procedure. A new Chebyshev collocation method is employed to approximate the porosity distribution and remapping is used to optimize memory usage. The adaptive approximation of the porosity distribution leads to a reduction of computational time and memory usage of up to two orders of magnitude. Moreover, the proposed model affords consistent performance: changing the nucleation pressure probability density function and/or the applied strain rate does not reduce accuracy or computational efficiency of the material model. The numerical performance of the model and algorithms presented is tested against three problems for high density tantalum: single void, one-dimensional uniaxial strain, and two-dimensional plate impact. Here, the results using the integration and algorithmic advances suggest a significant improvement in computational efficiency and accuracy over previous treatments for dynamic loading conditions.

42 ENGINEERING↗

Quantifying data retention of perpendicular spin-transfer-torque magnetic random access memory chips using an effective thermal stability factor method

The thermal stability of perpendicular Spin-Transfer-Torque Magnetic Random Access Memory (STT-MRAM) devices is investigated at chip level. Experimental data are analyzed in the framework of the Néel-Brown model including distributions of the thermal stability factor Δ. We show that in the low error rate regime important for applications, the effect of distributions of Δ can be described by a single quantity, the effective thermal stability factor Δ{sub eff}, which encompasses both the median and the standard deviation of the distributions. Data retention of memory chips can be assessed accurately by measuring Δ{sub eff} as a function of device diameter and temperature. We apply this method to show that 54 nm devices based on our perpendicular STT-MRAM design meet our 10 year data retention target up to 120 °C.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Eternal inflation, bubble collisions, and the disintegration of the persistence of memory

We compute the probability distribution for bubble collisions in an inflating false vacuum which decays by bubble nucleation. Our analysis generalizes previous work of Guth, Garriga, and Vilenkin to the case of general cosmological evolution inside the bubble, and takes into account the dynamics of the domain walls that form between the colliding bubbles. We find that incorporating these effects changes the results dramatically: the total expected number of bubble collisions in the past lightcone of a typical observer is N ∼ γ V{sub f}/V{sub i} , where γ is the fastest decay rate of the false vacuum, V{sub f} is its vacuum energy, and V{sub i} is the vacuum energy during inflation inside the bubble. This number can be large in realistic models without tuning. In addition, we calculate the angular position and size distribution of the collisions on the cosmic microwave background sky, and demonstrate that the number of bubbles of observable angular size is N{sub LS} ∼ (Ω{sub k}){sup 1/2}N, where Ω{sub k} is the curvature contribution to the total density at the time of observation. The distribution is almost exactly isotropic.

79 ASTRONOMY AND ASTROPHYSICS↗

Distributed Tomographic Reconstruction with Quantization

Conventional tomographic reconstruction typically depends on centralized servers for both data storage and computation, leading to concerns about memory limitations and data privacy. Distributed reconstruction algorithms mitigate these issues by partitioning data across multiple nodes, reducing server load and enhancing privacy. However, these algorithms often encounter challenges related to memory constraints and communication overhead between nodes. In this paper, we introduce a decentralized Alternating Directions Method of Multipliers (ADMM) with configurable quantization. By distributing local objectives across nodes, our approach is highly scalable and can efficiently reconstruct images while adapting to available resources. To overcome communication bottlenecks, we propose two quantization techniques based on K-means clustering and JPEG compression. Numerical experiments with benchmark images illustrate the tradeoffs between communication efficiency, memory use, and reconstruction accuracy.

Miao, Runxuan↗

TuckerMPI: A Parallel C++/MPI Software Package for Large-scale Data Compression via the Tucker Tensor Decomposition

With this study, our goal is compression of massive-scale grid-structured data, such as the multi-terabyte output of a high-fidelity computational simulation. For such data sets, we have developed a new software package called TuckerMPI, a parallel C++/MPI software package for compressing distributed data. The approach is based on treating the data as a tensor, i.e., a multidimensional array, and computing its truncated Tucker decomposition, a higher-order analogue to the truncated singular value decomposition of a matrix. The result is a low-rank approximation of the original tensor-structured data. Compression efficiency is achieved by detecting latent global structure within the data, which we contrast to most compression methods that are focused on local structure. In this work, we describe TuckerMPI, our implementation of the truncated Tucker decomposition, including details of the data distribution and in-memory layouts, the parallel and serial implementations of the key kernels, and analysis of the storage, communication, and computational costs. We test the software on 4.5 and 6.7 terabyte data sets distributed across 100 s of nodes (1,000 s of MPI processes), achieving compression ratios between 100 and 200,000×, which equates to 99--99.999% compression (depending on the desired accuracy) in substantially less time than it would take to even read the same dataset from a parallel file system. Moreover, we show that our method also allows for reconstruction of partial or down-sampled data on a single node, without a parallel computer so long as the reconstructed portion is small enough to fit on a single machine, e.g., in the instance of reconstructing/visualizing a single down-sampled time step or computing summary statistics. The code is available at https://gitlab.com/tensors/TuckerMPI.

97 MATHEMATICS AND COMPUTING↗

Vacancy structures and melting behavior in rock-salt GeSbTe

Ge-Sb-Te alloys have been widely used in optical/electrical memory storage. Because of the extremely fast crystalline-amorphous transition, they are also expected to play a vital role in next generation nonvolatile microelectronic memory devices. However, the distribution and structural properties of vacancies have been one of the key issues in determining the speed of melting (or amorphization), phase-stability, and heat-dissipation of rock-salt GeSbTe, which is crucial for its technological breakthrough in memory devices. Using spherical aberration-aberration corrected scanning transmission electron microscopy and atomic scale energy-dispersive X-ray mapping, we observe a new rock-salt structure with high-degree vacancy ordering (or layered-like ordering) at an elevated temperature, which is a result of phase transition from the rock-salt phase with randomly distributed vacancies. First-principles calculations reveal that the phase transition is an energetically favored process. Furthermore, molecular dynamics studies suggest that the melting of the cubic rock-salt phases is initiated at the vacancies, which propagate to nearby regions. The observation of multi-rock-salt phases suggests another route for multi-level data storage using GeSbTe.

36 MATERIALS SCIENCE↗

Kinetic energy classification and smoothing for compact B-spline basis sets in quantum Monte Carlo

Quantum Monte Carlo calculations of defect properties of transition metal oxides have become feasible in recent years due to increases in computing power. As the system size has grown, availability of on-node memory has become a limiting factor. Saving memory while minimizing computational cost is now a priority. The main growth in memory demand stems from the B-spline representation of the single particle orbitals, especially for heavier elements such as transition metals where semi-core states are present. Despite the associated memory costs, splines are computationally efficient. In this paper, we explore alternatives to reduce the memory usage of splined orbitals without significantly affecting numerical fidelity or computational efficiency. We make use of the kinetic energy operator to both classify and smooth the occupied set of orbitals prior to splining. By using a partitioning scheme based on the per-orbital kinetic energy distributions, we show that memory savings of about 50% is possible for select transition metal oxide systems. Finally, for production supercells of practical interest, our scheme incurs a performance penalty of less than 5%.

97 MATHEMATICS AND COMPUTING↗

Efficient Machine Learning Approach for Optimizing Scientific Computing Applications on Emerging HPC Architectures

Efficient parallel implementations of scientific applications on multi-core CPUs with accelerators such as GPUs and Xeon Phis is challenging. This requires - exploiting the data parallel architecture of the accelerator along with the vector pipelines of modern x86 CPU architectures, load balancing, and efficient memory transfer between different devices. It is relatively easy to meet these requirements for highly structured scientific applications. In contrast, a number of scientific and engineering applications are unstructured. Getting performance on accelerators for these applications is extremely challenging because many of these applications employ irregular algorithms which exhibit data-dependent control-ow and irregular memory accesses. Furthermore, these applications are often iterative with dependency between steps, and thus making it hard to parallelize across steps. As a result, parallelism in these applications is often limited to a single step. Numerical simulation of charged particles beam dynamics is one such application where the distribution of work and memory access pattern at each time step is irregular. Applications with these properties tend to present significant branch and memory divergence, load imbalance between different processor cores, and poor compute and memory utilization. Prior research on parallelizing such irregular applications have been focused around optimizing the irregular, data-dependent memory accesses and control-ow during a single step of the application independent of the other steps, with the assumption that these patterns are completely unpredictable. We observed that the structure of computation leading to control-ow divergence and irregular memory accesses in one step is similar to that in the next step. It is possible to predict this structure in the current step by observing the computation structure of previous steps. In this dissertation, we present novel machine learning based optimization techniques to address the parallel implementation challenges of such irregular applications on different HPC architectures. In particular, we use supervised learning to predict the computation structure and use it to address the control-ow and memory access irregularities in the parallel implementation of such applications on GPUs, Xeon Phis, and heterogeneous architectures composed of multi-core CPUs with GPUs or Xeon Phis. We use numerical simulation of charged particles beam dynamics simulation as a motivating example throughout the dissertation to present our new approach, though they should be equally applicable to a wide range of irregular applications. The machine learning approach presented here use predictive analytics and forecasting techniques to adaptively model and track the irregular memory access pattern at each time step of the simulation to anticipate the future memory access pattern. Access pattern forecasts can then be used to formulate optimization decisions during application execution which improves the performance of the application at a future time step based on the observations from earlier time steps. In heterogeneous architectures, forecasts can also be used to improve the memory performance and resource utilization of all the processing units to deliver a good aggregate performance. We used these optimization techniques and anticipation strategy to design a cache-aware, memory efficient parallel algorithm to address the irregularities in the parallel implementation of charged particles beam dynamics simulation on different HPC architectures. Experimental result using a diverse mix of HPC architectures shows that our approach in using anticipation strategy is effective in maximizing data reuse, ensuring workload balance, minimizing branch and memory divergence, and in improving resource utilization.

Arumugam, Kamesh↗

Hashing strategies for the Cray XMT.

Two of the most commonly used hashing strategies - linear probing and hashing with chaining - are adapted for efficient execution on a Cray XMT. These strategies are designed to minimize memory contention. Datasets that follow a power law distribution cause significant performance challenges to shared memory parallel hashing implementations. Experimental results show good scalability up to 128 processors on two power law datasets with different data types: integer and string. These implementations can be used in a wide range of applications.

99 GENERAL AND MISCELLANEOUS↗