Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “DATA PROCESSOR”

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 253 records · Page 14

A Pervasive Parallel Processing Framework for Data Visualization and Analysis at Extreme Scale

Most of today’s visualization libraries and applications are based off of what is known today as the visualization pipeline. In the visualization pipeline model, algorithms are encapsulated as “filtering” components with inputs and outputs. These components can be combined by connecting the outputs of one filter to the inputs of another filter. The visualization pipeline model is popular because it provides a convenient abstraction that allows users to combine algorithms in powerful ways. Unfortunately, the visualization pipeline cannot run effectively on exascale computers. Experts agree that the exascale machine will comprise processors that contain many cores. Furthermore, physical limitations will prevent data movement in and out of the chip (that is, between main memory and the processing cores) from keeping pace with improvements in overall compute performance. To use these processors to their fullest capability, it is essential to carefully consider memory access. This is where the visualization pipeline fails. Each filtering component in the visualization library is expected to take a data set in its entirety, perform some computation across all of the elements, and output the complete results. The process of iterating over all elements must be repeated in each filter, which is one of the worst possible ways to traverse memory when trying to maximize the number of executions per memory access. This project investigates a new type of visualization framework that exhibits a pervasive parallelism necessary to run on exascale machines. Our framework achieves this by defining algorithms in terms of functors, which are localized, stateless operations. Functors can be composited in much the same way as filters in the visualization pipeline. But, functors’ design allows them to be concurrently running on massive amounts of lightweight threads. Only with such fine-grained parallelism can we hope to fill the billions of threads we expect will be necessary for efficient computation on an exascale computer. This project concludes with a functional prototype containing pervasively parallel algorithms that perform demonstratively well on many-core processors. These algorithms are fundamental for performing data analysis and visualization at extreme scale.

97 MATHEMATICS AND COMPUTING↗

Extreme-Scale Algorithms & Software Resilience (EASIR) Architecture-Aware Algorithms for Scalable Performance and Resilience on Heterogeneous Architectures

This project addresses both communication-avoiding algorithms, and reproducible floating-point computation. Communication, i.e. moving data, either between levels of memory or processors over a network, is much more expensive per operation than arithmetic (measured in time or energy), so we seek algorithms that greatly reduce communication. We developed many new algorithms for both dense and sparse, and both direct and iterative linear algebra, attaining new communication lower bounds, and getting large speedups in many cases. We also extended this work in several ways: (1) We minimize writes separately from reads, since writes may be much more expensive than reads on emerging memory technologies, like Flash, sometimes doing asymptotically fewer writes than reads. (2) We extend the lower bounds and optimal algorithms to arbitrary algorithms that may be expressed as perfectly nested loops accessing arrays, where the array subscripts may be arbitrary affine functions of the loop indices (eg A(i), B(i,j+k, k+3*m-7, …) etc.). (3) We extend our communication-avoiding approach to some machine learning algorithms, such as support vector machines. This work has won a number of awards. We also address reproducible floating-point computation. We define reproducibility to mean getting bitwise identical results from multiple runs of the same program, perhaps with different hardware resources or other changes that should ideally not change the answer. Many users depend on reproducibility for debugging or correctness. However, dynamic scheduling of parallel computing resources, combined with nonassociativity of floating point addition, makes attaining reproducibility a challenge even for simple operations like summing a vector of numbers, or more complicated operations like the Basic Linear Algebra Subprograms (BLAS). We describe an algorithm that computes a reproducible sum of floating point numbers, independent of the order of summation. The algorithm depends only on a subset of the IEEE Floating Point Standard 754-2008, uses just 6 words to represent a “reproducible accumulator,” and requires just one read-only pass over the data, or one reduction in parallel. New instructions based on this work are being considered for inclusion in the future IEEE 754-2018 floating-point standard, and new reproducible BLAS are being considered for the next version of the BLAS standard.

97 MATHEMATICS AND COMPUTING↗

MPI-FAUN: An MPI-Based Framework for Alternating-Updating Nonnegative Matrix Factorization

Non-negative matrix factorization (NMF) is the problem of determining two non-negative low rank factors W and H, for the given input matrix A, such that A≈WH. NMF is a useful tool for many applications in different domains such as topic modeling in text mining, background separation in video analysis, and community detection in social networks. Despite its popularity in the data mining community, there is a lack of efficient parallel algorithms to solve the problem for big data sets. The main contribution of this work is a new, high-performance parallel computational framework for a broad class of NMF algorithms that iteratively solves alternating non-negative least squares (NLS) subproblems for W and H. It maintains the data and factor matrices in memory (distributed across processors), uses MPI for interprocessor communication, and, in the dense case, provably minimizes communication costs (under mild assumptions). The framework is flexible and able to leverage a variety of NMF and NLS algorithms, including Multiplicative Update, Hierarchical Alternating Least Squares, and Block Principal Pivoting. Our implementation allows us to benchmark and compare different algorithms on massive dense and sparse data matrices of size that spans from few hundreds of millions to billions. We demonstrate the scalability of our algorithm and compare it with baseline implementations, showing significant performance improvements. The code and the datasets used for conducting the experiments are available online.

2D↗

Self-adaptive sampling rate data acquisition in JET's correlation reflectometer

Data acquisition systems with self-adaptive sampling rate capabilities have been proposed as a solution to reduce the shear amount of data collected in every discharge of present fusion devices. This paper discusses the design of such a system for its use in the KG8B correlation reflectometer at JET. The system, which is based on the ITMS platform, continuously adapts the sample rate during the acquisition depending on the signal bandwidth. Data are acquired continuously at the expected maximum sample rate and transferred to a memory buffer in the host processor. Thereafter the rest of the process is based on software. Data are read from the memory buffer in blocks and for each block an intelligent decimation algorithm is applied. The decimation algorithm determines the signal bandwidth for each block in order to choose the optimum sample rate for that block, and from there the decimation factor to be used. Memory buffers are used to adapt the throughput of the three main software modules (data acquisition, processing, and storage) following a typical producer-consumer architecture. The system optimizes the amount of data collected while maintaining the same information. Design issues are discussed and results of performance evaluation are presented.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Enhancements to the RAVEN code in FY16

The RAVEN code has been under development at the Idaho National Laboratory since 2012. Its main goal is to create a multi-purpose platform for the deploying of all the capabilities needed for Probabilistic Risk Assessment, uncertainty quantification, data mining analysis and optimization studies. RAVEN has currently reached a good level of maturity in terms of deployed state-of-art and advanced capabilities. The main subject of this report is to show the activities that have been recently accomplished: • Implementation of ensemble modeling for time-series, and • initial implementation of model validation for surrogate models, and • advanced visualization capability for topology based data analysis The development of ensemble modeling for time-series has been performed in order to begin tackling the needs of those RISMC applications that need to communicate 1-Dimensional information (e.g. power histories, etc.) among different models. In this document the implementation details and an application example is reported. The second subject of this report is about the initial development of methods, within the RAVEN framework, to assess the validity of the predictive capabilities of surrogate models. Indeed, after the construction of a surrogate tight to a certain physical model, it is crucial to assess the goodness of its representation, in order to be confident with its prediction. In this initial activity, a cross-validation technique has been employed. This report will highlight the implementation details and proof its correct implementation by an application example. The final subject of this report is about the implementation of advanced visualization capability in RAVEN, for interactive data analysis. Indeed, RAVEN offers several post-processing capabilities that can structurally decompose data extracted from experimental results offering both data clustering/partitioning and dimensionality reduction techniques. A disadvantage of the workflow available in RAVEN is that it treats these as black box operations and the user is expected to know specific information about their data including the number of partitions to expect in some cases or the “correct” parameter settings for a particular algorithm. In order to overcome this limitation, it has been added an interactive user interface that can be run in RAVEN to explore different parameter settings on the fly specifically for the topological post-processor and allows the user to explore the data interactively. The design of this user interface is generalizable to other algorithms and can be used as a model to generate more dynamic and user-friendly visualization capabilities within RAVEN.

97 MATHEMATICS AND COMPUTING↗

OpenMP in VASP: Threading and SIMD

Abstract The Vienna Ab initio Simulation Package (VASP) is a widely used electronic structure code that originally exploits process‐level parallelism through the Message Passing Interface (MPI) for work distribution within and across nodes. Architectural changes of modern parallel processors urge programmers to address thread‐ and data‐level parallelism as well to benefit most from the available compute resources within a node. We describe for VASP how to approach for an MPI + OpenMP parallelization including data‐level parallelism through OpenMP SIMD constructs together with a generic high‐level vector coding scheme. We can demonstrate an improved scalability of VASP and more than 20% gain over the MPI‐only version as well as a 2× increased performance of collective operations using the multiple‐endpoint MPI feature. The high‐level vector coding scheme applied to VASP's general gradient approximation routine gives up 9× performance gain on AVX512 platforms with the Intel compiler.

Wende, Florian↗

Quantum-assisted associative adversarial network: applying quantum annealing in deep learning

Abstract Generative models have the capacity to model and generate new examples from a dataset and have an increasingly diverse set of applications driven by commercial and academic interest. In this work, we present an algorithm for learning a latent variable generative model via generative adversarial learning where the canonical uniform noise input is replaced by samples from a graphical model. This graphical model is learned by a Boltzmann machine which learns low-dimensional feature representation of data extracted by the discriminator. A quantum processor can be used to sample from the model to train the Boltzmann machine. This novel hybrid quantum-classical algorithm joins a growing family of algorithms that use a quantum processor sampling subroutine in deep learning, and provides a scalable framework to test the advantages of quantum-assisted learning. For the latent space model, fully connected, symmetric bipartite and Chimera graph topologies are compared on a reduced stochastically binarized MNIST dataset, for both classical and quantum sampling methods. The quantum-assisted associative adversarial network successfully learns a generative model of the MNIST dataset for all topologies. Evaluated using the Fréchet inception distance and inception score, the quantum and classical versions of the algorithm are found to have equivalent performance for learning an implicit generative model of the MNIST dataset. Classical sampling is used to demonstrate the algorithm on the LSUN bedrooms dataset, indicating scalability to larger and color datasets. Though the quantum processor used here is a quantum annealer, the algorithm is general enough such that any quantum processor, such as gate model quantum computers, may be substituted as a sampler.

Wilson, Max (ORCID:0000000207983391)↗

Nanoscale Probing of Electrical Memory Effects in van der Waals Layered PdSe 2

Tunable electronic materials that can be switched between different impedance states are fundamental to the hardware elements for neuromorphic computing architectures. This “brain-like” computing paradigm uses highly paralleled and colocated data processing, leading to greatly improved energy efficiency and performance compared to traditional architectures in which data have to be frequently transferred between processor and memory. In this work, we use scanning microwave impedance microscopy for nanoscale electrical and electronic characterization of two-dimensional layered semiconductor PdSe 2 to probe neuromorphic properties. The local resolution of tens of nanometers reveals significant differences in electronic behavior between and within PdSe 2 nanosheets (NSs). In particular, we detected both n-type and p-type behaviors, although previous reports only point to ambipolar n-type dominating characteristics. Nanoscale capacitance–voltage curves and subsequent calculation of characteristic maps revealed a hysteretic behavior originating from the creation and erasure of Se vacancies as well as the switching of defect charge states. In addition, stacks consisting of two NSs show enhanced resistive and capacitive switching, which is attributed to trapped charge carriers at the interfaces between the stacked NSs. Stacking n- and p-type NSs results in a combined behavior that allows one to tune electrical characteristics. In conclusion, as local inhomogeneities of electrical and electronic behavior can have a significant impact on the overall device performance, the demonstrated nanoscale characterization and analysis will be applicable to a wide range of semiconducting materials.

2D material↗

Relaxation of stationary states on a quantum computer yields a unique spectroscopic fingerprint of the computer’s noise

Abstract Quantum computing has the potential to revolutionize computing, but its significant sensitivity to noise requires sophisticated error correction and mitigation. Traditionally, noise on the quantum device is characterized directly through qubit and gate measurements, but this approach has drawbacks in that it does not adequately capture the effect of noise on realistic multi-qubit applications. In this paper, we simulate the relaxation of stationary quantum states on a quantum computer to obtain a unique spectroscopic fingerprint of the computer’s noise. In contrast to traditional approaches, we obtain the frequency profile of the noise as it is experienced by the simulated stationary quantum states. Data from multiple superconducting-qubit IBM processors show that noise generates a bath within the simulation that exhibits both colored noise and non-Markovian behavior. Our results provide a direction for noise mitigation but also suggest how to use noise for quantum simulations of open systems.

97 MATHEMATICS AND COMPUTING↗

Scaling Deep Learning on GPU and Knights Landing clusters

The speed of deep neural networks training has become a big bottleneck of deep learning research and development. For example, training GoogleNet by ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training process, the current deep learning systems heavily rely on the hardware accelerators. However, these accelerators have limited on-chip memory compared with CPUs. To handle large datasets, they need to fetch data from either CPU memory or remote processors. We use both self-hosted Intel Knights Landing (KNL) clusters and multi-GPU clusters as our target platforms. From an algorithm aspect, current distributed machine learning systems are mainly designed for cloud systems. These methods are asynchronous because of the slow network and high fault-tolerance requirement on cloud systems. We focus on Elastic Averaging SGD (EASGD) to design algorithms for HPC clusters. Original EASGD used round-robin method for communication and updating. The communication is ordered by the machine rank ID, which is inefficient on HPC clusters. First, we redesign four efficient algorithms for HPC systems to improve EASGD's poor scaling on clusters. Async EASGD, Async MEASGD, and Hogwild EASGD are faster \textcolor{black}{than} their existing counterparts (Async SGD, Async MSGD, and Hogwild SGD, resp.) in all the comparisons. Finally, we design Sync EASGD, which ties for the best performance among all the methods while being deterministic. In addition to the algorithmic improvements, we use some system-algorithm codesign techniques to scale up the algorithms. By reducing the percentage of communication from 87% to 14%, our Sync EASGD achieves 5.3x speedup over original EASGD on the same platform. We get 91.5% weak scaling efficiency on 4253 KNL cores, which is higher than the state-of-the-art implementation.

60 APPLIED LIFE SCIENCES↗

xSim: The Extreme-Scale Simulator

Investigating parallel application performance properties at scale is becoming an important part of high-performance computing (HPC) application development and deployment. The Extreme-scale Simulator (xSim) is a performance investigation toolkit that permits running an application in a controlled environment at extreme scale without the need for a respective extreme-scale HPC system. Using a lightweight parallel discrete event simulation, xSim executes a parallel application with a virtual wall clock time, such that performance data can be extracted based on a processor model and a network model. This paper presents significant enhancements to the xSim toolkit prototype that provide a more complete Message Passing Interface (MPI) support and improve its versatility. These enhancements include full virtual MPI group, communicator and collective communication support, and global variables support. The new capabilities are demonstrated by executing the entire NAS Parallel Benchmark suite in a simulated HPC environment.

99 GENERAL AND MISCELLANEOUS↗

Three-dimensional broadband nonlinear phased array imaging

System include an ultrasonic transducer configured to couple to a nondestructive testing (NDT) sample and configured to produce and direct an ultrasonic probe wave at a selected frequency into a subsurface region of the NDT sample, a 3D laser scanning vibrometer configured to direct a detection beam in a scan area on a surface of the NDT sample and to receive a return beam from the scan area, and to detect, based on the return beam, a 3D motion of the surface across a wideband frequency range, and a processor, and a memory configured with instructions that, when executed by the processor, cause the processor to produce sub-surface image data of the NDT sample at multiple harmonics of the selected frequency in the wideband frequency range based on the detected 3D surface motion, wherein the sub-surface image data describes a nonlinear defect response produced in the NDT sample by interaction of the ultrasonic probe wave with the subsurface region.

Remillieux, Marcel↗

Lanthanum halide scintillators for time-of-flight 3-D pet

A Lanthanum Halide scintillator (for example LaCl.sub.3 and LaBr.sub.3) with fast decay time and good timing resolution, as well as high light output and good energy resolution, is used in the design of a PET scanner. The PET scanner includes a cavity for accepting a patient and a plurality of PET detector modules arranged in an approximately cylindrical configuration about the cavity. Each PET detector includes a Lanthanum Halide scintillator having a plurality of Lanthanum Halide crystals, a light guide, and a plurality of photomultiplier tubes arranged respectively peripherally around the cavity. The good timing resolution enables a time-of-flight (TOF) PET scanner to be developed that exhibits a reduction in noise propagation during image reconstruction and a gain in the signal-to-noise ratio. Such a PET scanner includes a time stamp circuit that records the time of receipt of gamma rays by respective PET detectors and provides timing data outputs that are provided to a processor that, in turn, calculates time-of-flight (TOF) of gamma rays through a patient in the cavity and uses the TOF of gamma rays in the reconstruction of images of the patient.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Field Programmable Gate Array Data Capture for Control Systems

Some Industrial Control Systems (ICS) networks are based on protocols such as Serial and Industrial Ethernet. These protocols currently have no existing cybersecurity monitoring tools, leaving a large gap in the cyber defense of critical infrastructure. In order to analyze such ICS traffic, it is first necessary to implement methods of capturing the ICS data. Whereas traditional methods of analyzing data would use microprocessors, the nature of high-speed analog data can be difficult to implement on such a versatile processor, as they are rather inefficient for doing a single task. Whereas Field Programmable Gate Arrays (FPGAs) provide an adequate tool in analyzing high speed data, as despite the lack of program versatility, Programmable Logic can implement a solution with minimal clock cycles, allowing time for each new packet of data to be captured before a new data sample is taken.

42 ENGINEERING↗

Bin-Hash Indexing: A Parallel Method for Fast Query Processing

This paper presents a new parallel indexing data structure for answering queries. The index, called Bin-Hash, offers extremely high levels of concurrency, and is therefore well-suited for the emerging commodity of parallel processors, such as multi-cores, cell processors, and general purpose graphics processing units (GPU). The Bin-Hash approach first bins the base data, and then partitions and separately stores the values in each bin as a perfect spatial hash table. To answer a query, we first determine whether or not a record satisfies the query conditions based on the bin boundaries. For the bins with records that can not be resolved, we examine the spatial hash tables. The procedures for examining the bin numbers and the spatial hash tables offer the maximum possible level of concurrency; all records are able to be evaluated by our procedure independently in parallel. Additionally, our Bin-Hash procedures access much smaller amounts of data than similar parallel methods, such as the projection index. This smaller data footprint is critical for certain parallel processors, like GPUs, where memory resources are limited. To demonstrate the effectiveness of Bin-Hash, we implement it on a GPU using the data-parallel programming language CUDA. The concurrency offered by the Bin-Hash index allows us to fully utilize the GPU's massive parallelism in our work; over 12,000 records can be simultaneously evaluated at any one time. We show that our new query processing method is an order of magnitude faster than current state-of-the-art CPU-based indexing technologies. Additionally, we compare our performance to existing GPU-based projection index strategies.

97 MATHEMATICS AND COMPUTING↗

Thought Leaders during Crises in Massive Social Networks

The vast amount of social media data that can be gathered from the internet coupled with workflows that utilize both commodity systems and massively parallel supercomputers, such as the Cray XMT, open new vistas for research to support health, defense, and national security. Computer technology now enables the analysis of graph structures containing more than 4 billion vertices joined by 34 billion edges along with metrics and massively parallel algorithms that exhibit near-linear scalability according to number of processors. The challenge lies in making this massive data and analysis comprehensible to an analyst and end-users that require actionable knowledge to carry out their duties. Simply stated, we have developed language and content agnostic techniques to reduce large graphs built from vast media corpora into forms people can understand. Specifically, our tools and metrics act as a survey tool to identify thought leaders' -- those members that lead or reflect the thoughts and opinions of an online community, independent of the source language.

97 MATHEMATICS AND COMPUTING↗

Scalable load-balance measurement for SPMD codes

Good load balance is crucial on very large parallel systems, but the most sophisticated algorithms introduce dynamic imbalances through adaptation in domain decomposition or use of adaptive solvers. To observe and diagnose imbalance, developers need system-wide, temporally-ordered measurements from full-scale runs. This potentially requires data collection from multiple code regions on all processors over the entire execution. Doing this instrumentation naively can, in combination with the application itself, exceed available I/O bandwidth and storage capacity, and can induce severe behavioral perturbations. We present and evaluate a novel technique for scalable, low-error load balance measurement. This uses a parallel wavelet transform and other parallel encoding methods. We show that our technique collects and reconstructs system-wide measurements with low error. Compression time scales sublinearly with system size and data volume is several orders of magnitude smaller than the raw data. The overhead is low enough for online use in a production environment.

99 GENERAL AND MISCELLANEOUS↗

Systems and methods for tensor scheduling

A technique for efficient scheduling of operations in a program for parallelized execution thereof using a multi-processor runtime environment having two or more processors includes constraining the type or number of loop optimization transforms that may be explored such that memory and processing capacity available for the scheduling task are not exceeded, while facilitating a tradeoff between memory locality, parallelization, and/or data communication between memory modules of the multi-processor runtime environment.

Meister, Benoit J.↗