Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Multiple processors”

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 163 records · Page 9

Load Balancing Scientific Applications

The largest supercomputers have millions of independent processors, and concurrency levels are rapidly increasing. For ideal efficiency, developers of the simulations that run on these machines must ensure that computational work is evenly balanced among processors. Assigning work evenly is challenging because many large modern parallel codes simulate behavior of physical systems that evolve over time, and their workloads change over time. Furthermore, the cost of imbalanced load increases with scale because most large-scale scientific simulations today use a Single Program Multiple Data (SPMD) parallel programming model, and an increasing number of processors will wait for the slowest one at the synchronization points. To address load imbalance, many large-scale parallel applications use dynamic load balance algorithms to redistribute work evenly. The research objective of this dissertation is to develop methods to decide when and how to load balance the application, and to balance it effectively and affordably. We measure and evaluate the computational load of the application, and develop strategies to decide when and how to correct the imbalance. Depending on the simulation, a fast, local load balance algorithm may be suitable, or a more sophisticated and expensive algorithm may be required. We developed a model for comparison of load balance algorithms for a specific state of the simulation that enables the selection of a balancing algorithm that will minimize overall runtime.

97 MATHEMATICS AND COMPUTING↗

Grid scheduling divisible loads from two sources

To date closed form solutions for optimal finish time and job allocation are largely obtained only for network topologies with a single load originating (root) processor. However in large-scale data intensive problems with geographically distributed resources, load is generated from multiple sources. This paper introduces a new divisible load scheduling strategy for single level tree networks with two load originating processors. Solutions for an optimal allocation of fractions of load to nodes in single level tree networks are obtained via linear programming. A unique scheduling strategy that allows one to obtain closed form solutions for the optimal finish time and load allocation for each processor in the network is also presented. The tradeoff between linear programming and closed form solutions in terms of underlying assumptions is examined. Finally, a performance evaluation of a two source homogeneous single level tree network with concurrent communication strategy is presented.

97 MATHEMATICS AND COMPUTING↗

Efficient Breadth-First Search on the Cell/BE Processor

Multi-core processors are a shift of paradigm in computer architecture that promises a dramatic increase in performance. But multi-core processors also bring an unprecedented level of complexity in algorithmic design and software development. In this paper we describe the challenges involved in designing a breadth-first search (BFS) for advanced multi-core processors. The proposed methodology combines a high-level algorithmic design that captures the machine-independent aspects, to guarantee portability with performance to future processors, with a low-level implementation that embeds processor-specific optimizations. Using a fine-graind global coordination strategy derived by the Bulk-Synchronous Parallel (BSP) model we have determined an accurate performance model that has guided the implementation and the optimization of our algorithms. To validate our approach, we use a state-of-the-art multicore processor, the Cell Broadband Engine (Cell BE). Our experiments, obtained on a pre-production Cell BE board running at 3.2 GHz, show almost linear speedups when using multiple synergistic processing elements, and an impressive level of performance when compared to other processors. The Cell BE is typically an order of magnitude faster than conventional processors, such as the AMD Opteron and the Intel Pentium 4 and Woodcrest, an order of magnitude faster than the MTA-2 multi-threaded processor, and two orders of magnitude faster than a BlueGene/L processor. Index Terms—Multi-core processors, Parallel Computing, Cell Broadband Engine, Parallelization Techniques, Graph Exploration Algorithms, Breadth-First Search, BFS.

97 MATHEMATICS AND COMPUTING↗

Solution to PDEs using radial basis function finite-differences (RBF-FD) on multiple GPUs

This paper presents parallelization strategies for the radial basis function-finite difference (RBF-FD) method. As a generalized finite differencing scheme, the RBF-FD method functions without the need for underlying meshes to structure nodes. It offers high-order accuracy approximation and scales as O(N) per time step, with N being with the total number of nodes. To our knowledge, this is the first implementation of the RBF-FD method to leverage GPU accelerators for the solution of PDEs. Additionally, this implementation is the first to span both multiple CPUs and multiple GPUs. OpenCL kernels target the GPUs and inter-processor communication and synchronization is managed by the Message Passing Interface (MPI). We verify our implementation of the RBF-FD method with two hyperbolic PDEs on the sphere, and demonstrate up to 9x speedup on a commodity GPU with unoptimized kernel implementations. On a high performance cluster, the method achieves up to 7x speedup for the maximum problem size of 27,556 nodes.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

GPU-based parallel algorithm for generating massive scale-free networks using the preferential attachment model

A novel parallel algorithm is presented for generating random scale-free networks using the preferential-attachment model. The algorithm, named cuPPA, is custom-designed for single instruction multiple data (SIMD) style of parallel processing supported by modern processors such as graphical processing units (GPUs). To the best of our knowledge, our algorithm is the first to exploit GPUs, and also the fastest implementation available today, to generate scale-free networks using the preferential attachment model. A detailed performance study is presented to understand the scalability and runtime characteristics of the cuPPA algorithm. In one of the best cases, when executed on an NVidia GeForce 1080 GPU, cuPPA generates a scale-free network of two billion edges in less than 3 seconds.

Alam, Maksudul↗

Spatial Locality-Aware Cache Partitioning for Effective Cache Sharing, In: 2015 44th International Conference on Parallel Processing

In modern multi-core processors, last-level caches (LLCs) are typically shared among multiple cores. Previous works have shown that such sharing is beneficial as different workloads have different needs for cache capacity, and logical partitioning of capacity can improve system performance. However, what is missing in previous works on partitioning shared LLCs is that the heterogeneity in spatial locality among workloads has not been explored. In other words, all the cores use the same block/line size in shared LLCs. In this work, we highlight that exploiting spatial locality enables much more effective cache sharing. The fundamental reason is that for many memory intensive workloads, their cache capacity requirements can be drastically reduced when a large block size is employed, therefore they can effectively donate more capacity to other workloads. To leverage spatial locality for cache partitioning effectively, we first propose a simple yet effective mechanism to measure both spatial and temporal locality at run-time. The locality information is then used to determine both the proper block size and the capacity assigned to each workload. Our experiments show that our Spatial Locality-aware Cache Partitioning (SLCP) significantly outperforms the previous works. We also present several case studies that dissect the effectiveness of SLCP compared to the existing approaches.

Computer Science↗

Generating Billion-Edge Scale-Free Networks in Seconds: Performance Study of a Novel GPU-based Preferential Attachment Model

A novel parallel algorithm is presented for generating random scale-free networks using the preferential-attachment model. The algorithm, named cuPPA, is custom-designed for single instruction multiple data (SIMD) style of parallel processing supported by modern processors such as graphical processing units (GPUs). To the best of our knowledge, our algorithm is the first to exploit GPUs, and also the fastest implementation available today, to generate scale free networks using the preferential attachment model. A detailed performance study is presented to understand the scalability and runtime characteristics of the cuPPA algorithm. In one of the best cases, when executed on an NVidia GeForce 1080 GPU, cuPPA generates a scale free network of a billion edges in less than 2 seconds.

97 MATHEMATICS AND COMPUTING↗

Deep Trek Re-configurable Processor for Data Acquisition (RPDA)

This report summarizes technical progress achieved during the cooperative research agreement between Honeywell and U.S. Department of Energy to develop a high-temperature Re-configurable Processor for Data Acquisition (RPDA). The RPDA development has incorporated multiple high-temperature (225C) electronic components within a compact co-fired ceramic Multi-Chip-Module (MCM) package. This assembly is suitable for use in down-hole oil and gas applications. The RPDA module is programmable to support a wide range of functionality. Specifically this project has demonstrated functional integrity of the RPDA package and internal components, as well as functional integrity of the RPDA configured to operate as a Multi-Channel Data Acquisition Controller. This report reviews the design considerations, electrical hardware design, MCM package design, considerations for manufacturing assembly, test and screening, and results from prototype assembly and characterization testing.

02 PETROLEUM↗

GPU-accelerated Software Library for Unsteady Flamelet Modeling of Turbulent Combustion with Complex Chemical Kinetics

Modeling and simulation of turbulent combustion are computationally demanding and challenging problems due to the complexity of turbulent flows and the complex chemical reaction kinetics. Recent advances in high performance computing (HPC) has allowed simulations to utilize increasing computer power to obtain newer insights that were not feasible before. The arrival of GPUs and many-core processors is set to revolutionize the HPC landscape. This paper presents an unsteady flamelet software library developed for utilizing GPUs as an onboard accelerator while the flow solver continues to run on conventional host processors using legacy software. The software was architected to expose multiple levels of parallelism and concurrency available in the unsteady flamelet model which is also necessary to realize the full potential of these accelerators. While the model and its coupling with the flow solver have been known for some time, the potential to target these models and libraries for newer HPC architectures provides a greater incentive for pursuing combustion model development.

Sankaran, Ramanan↗

On Undecidability Aspects of Resilient Computations and Implications to Exascale

Future Exascale computing systems with a large number of processors, memory elements and interconnection links, are expected to experience multiple, complex faults, which affect both applications and operating-runtime systems. A variety of algorithms, frameworks and tools are being proposed to realize and/or verify the resilience properties of computations that guarantee correct results on failure-prone computing systems. We analytically show that certain resilient computation problems in presence of general classes of faults are undecidable, that is, no algorithms exist for solving them. We first show that the membership verification in a generic set of resilient computations is undecidable. We describe classes of faults that can create infinite loops or non-halting computations, whose detection in general is undecidable. We then show certain resilient computation problems to be undecidable by using reductions from the loop detection and halting problems under two formulations, namely, an abstract programming language and Turing machines, respectively. These two reductions highlight different failure effects: the former represents program and data corruption, and the latter illustrates incorrect program execution. These results call for broad-based, well-characterized resilience approaches that complement purely computational solutions using methods such as hardware monitors, co-designs, and system- and application-specific diagnosis codes.

Exascale systems↗

PLATO (Parallel Load Assignment Tool): A Parallel Workload Partitioner for Particle-based Methods

A key first step in many large-scale high-performance scientific computing applications is the distribution of the computational domain across all participating processors of a parallel computing platform such as a shared memory machine, cluster or a tightly-coupled supercomputer. The primary goal of such a partitioning step is to ensure that the computational work performed by each process remains load-balanced throughout the execution of the parallel application. In this project, a general-purpose parallel partitioning tool, code-named PLATO, was designed and implemented. PLATO assumes that the workload is a set of points in an n-dimensional metric space. These points are abstractions of computational units that may represent point-like computational units such as celestial objects in astrophysics, atoms in materials sciences and molecules in biology/chemistry, or may represent more abstract computational units such as grid points in finite difference methods, grid elements in finite element methods or n-dimensional points in the feature space of a host of other applications. PLATO supports several key and useful capabilities, namely: (a) multiple load partitioning strategies, (b) support for arbitrary user-defined processor topology, (c) pre-fetching data from neighboring processors based on user-defined parameters, and (d) support for arbitrary number and types of attributes associated with each point in the computational domain. PLATO is implemented in the C programming language using Message Passing Interface (MPI) for inter-processor communications and its parallel performance tested on up to 48 processors of a shared memory machine. This report includes a detailed performance study of PLATO using parallel speedup and iso-efficiency as the two main metrics of parallel performance. In addition, a user manual and DOXYGEN generated code documentation have been included for completeness.

97 MATHEMATICS AND COMPUTING↗

Integration of Ag-CBRAM crossbars and Mott ReLU neurons for efficient implementation of deep neural networks in hardware

In-memory computing with emerging non-volatile memory devices (eNVMs) has shown promising results in accelerating matrix-vector multiplications. However, activation function calculations are still being implemented with general processors or large and complex neuron peripheral circuits. Here, we present the integration of Ag-based conductive bridge random access memory (Ag-CBRAM) crossbar arrays with Mott rectified linear unit (ReLU) activation neurons for scalable, energy and area-efficient hardware (HW) implementation of deep neural networks. We develop Ag-CBRAM devices that can achieve a high ON/OFF ratio and multi-level programmability. Compact and energy-efficient Mott ReLU neuron devices implementing ReLU activation function are directly connected to the columns of Ag-CBRAM crossbars to compute the output from the weighted sum current. We implement convolution filters and activations for VGG-16 using our integrated HW and demonstrate the successful generation of feature maps for CIFAR-10 images in HW. Our approach paves a new way toward building a highly compact and energy-efficient eNVMs-based in-memory computing system.

Mott insulators↗

DYNA3D/ParaDyn Regression Test Suite Inventory

The following table constitutes an initial assessment of feature coverage across the regression test suite used for DYNA3D and ParaDyn. It documents the regression test suite at the time of production release 10.1 in September 2010. The columns of the table represent groupings of functionalities, e.g., material models. Each problem in the test suite is represented by a row in the table. All features exercised by the problem are denoted by a check mark in the corresponding column. The definition of ''feature'' has not been subdivided to its smallest unit of user input, e.g., algorithmic parameters specific to a particular type of contact surface. This represents a judgment to provide code developers and users a reasonable impression of feature coverage without expanding the width of the table by several multiples. All regression testing is run in parallel, typically with eight processors. Many are strictly regression tests acting as a check that the codes continue to produce adequately repeatable results as development unfolds, compilers change and platforms are replaced. A subset of the tests represents true verification problems that have been checked against analytical or other benchmark solutions. Users are welcomed to submit documented problems for inclusion in the test suite, especially if they are heavily exercising, and dependent upon, features that are currently underrepresented.

42 ENGINEERING↗

DYNA3D/ParaDyn Regression Test Suite Inventory

The following table constitutes an initial assessment of feature coverage across the regression test suite used for DYNA3D and ParaDyn. It documents the regression test suite at the time of preliminary release 16.1 in September 2016. The columns of the table represent groupings of functionalities, e.g., material models. Each problem in the test suite is represented by a row in the table. All features exercised by the problem are denoted by a check mark (√) in the corresponding column. The definition of “feature” has not been subdivided to its smallest unit of user input, e.g., algorithmic parameters specific to a particular type of contact surface. This represents a judgment to provide code developers and users a reasonable impression of feature coverage without expanding the width of the table by several multiples. All regression testing is run in parallel, typically with eight processors, except problems involving features only available in serial mode. Many are strictly regression tests acting as a check that the codes continue to produce adequately repeatable results as development unfolds; compilers change and platforms are replaced. A subset of the tests represents true verification problems that have been checked against analytical or other benchmark solutions. Users are welcomed to submit documented problems for inclusion in the test suite, especially if they are heavily exercising, and dependent upon, features that are currently underrepresented.

97 MATHEMATICS AND COMPUTING↗

End-System Aware Peak Link Utilization Transport for Big Data Transfer over Dedicated Connections

High-performance networks provisioning 10/40-Gbps dedicated channels have been rapidly developed andincreasingly deployed to support big data transfers on the order of terabytes or petabytes in large-scale network-intensiveapplications. However, end users have not seen corresponding increase in application throughput mainly because traditional end-to-end transport methods are not optimized for such connections. New congestion or flow control mechanisms are desirable to meet the challenges brought to transport protocol design by dedicated connections. The advent and proliferation of multi-core processors make it now possible to improve application throughput by providing multiple processing and network resources to a single data transfer. Based on the existing PLUT method, we proposed a new dynamic buffer management scheme and an automatic parallelism tuning mechanism to achieve the maximal bottleneck rate. The performance superiority of PLUT is justified by the experimental results collected over a local 10Gbps dedicated connection.

Rao, Nageswara↗

Sparse matrix‐vector and matrix‐multivector products for the truncated SVD on graphics processors

Summary Many practical algorithms for numerical rank computations implement an iterative procedure that involves repeated multiplications of a vector, or a collection of vectors, with both a sparse matrix and its transpose. Unfortunately, the realization of these sparse products on current high performance libraries often deliver much lower arithmetic throughput when the matrix involved in the product is transposed. In this work, we propose a hybrid sparse matrix layout, named CSRC, that combines the flexibility of some well‐known sparse formats to offer a number of appealing properties: (1) CSRC can be obtained at low cost from the popular CSR (compressed sparse row) format; (2) CSRC has similar storage requirements as CSR; and especially, (3) the implementation of the sparse product kernels delivers high performance for both the direct product and its transposed variant on modern graphics accelerators thanks to a significant reduction of atomic operations compared to a conventional implementation based on CSR. This solution thus renders considerably higher performance when integrated into an iterative algorithm for the truncated singular value decomposition (SVD), such as the randomized SVD or, as demonstrated in the experimental results, the block Golub–Kahan–Lanczos algorithm.

Aliaga, José I.↗

Method and apparatus to debug an integrated circuit chip via synchronous clock stop and scan

An apparatus and method for evaluating a state of an electronic or integrated circuit (IC), each IC including one or more processor elements for controlling operations of IC sub-units, and each the IC supporting multiple frequency clock domains. The method comprises: generating a synchronized set of enable signals in correspondence with one or more IC sub-units for starting operation of one or more IC sub-units according to a determined timing configuration; counting, in response to one signal of the synchronized set of enable signals, a number of main processor IC clock cycles; and, upon attaining a desired clock cycle number, generating a stop signal for each unique frequency clock domain to synchronously stop a functional clock for each respective frequency clock domain; and, upon synchronously stopping all on-chip functional clocks on all frequency clock domains in a deterministic fashion, scanning out data values at a desired IC chip state. The apparatus and methodology enables construction of a cycle-by-cycle view of any part of the state of a running IC chip, using a combination of on-chip circuitry and software.

47 OTHER INSTRUMENTATION↗

Fully Arbitrary Control of Frequency-Bin Qubits

Accurate control of two-level systems is a longstanding problem in quantum mechanics. One such quantum system is the frequency-bin qubit: a single photon existing in superposition of two discrete frequency modes. In this Letter, we demonstrate fully arbitrary control of frequency-bin qubits in a quantum frequency processor for the first time. Furthermore, we numerically establish optimal settings for multiple configurations of electro-optic phase modulators and pulse shapers, experimentally confirming near-unity mode-transformation fidelity for all fundamental rotations. Performance at the single-photon level is validated through the rotation of a single frequency-bin qubit to 41 points spread over the entire Bloch sphere, as well as tracking of the state path followed by the output of a tunable frequency beam splitter, with Bayesian tomography confirming state fidelities Fρ>0.98 for all cases. Overall, such high-fidelity transformations expand the practical potential of frequency encoding in quantum communications, offering exceptional precision and low noise in general qubit manipulation.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗