Search NASA⌕ Search

SEARCH · Search NASA

Results for “Computer hardware”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,099 records · Page 61

Overlapping Computation and Communication: Barrier Algorithms and ConnectX-2 CORE-Direct Capabilities

This paper explores the computation and communication overlap capabilities enabled by the new CORE-Direct hardware capabilities introduced in the InfiniBand (IB) Host Channel Adapter (HCA) ConnectX-2. These capabilities enable the progression and completion of data-dependent communications sequences to progress and complete at the network level without any Central Processing Unit (CPU) involvement. We use the latency dominated nonblocking barrier algorithm in this study, and find that at 64 process count, a contiguous time slot of about 80 percent of the nonblocking barrier time is available for computation. This time slot increases as the number of processes participating increases. In contrast, CPU based implementations provide a time slot of up to 30 percent of the nonblocking barrier time. This bodes well for the scalability of simulations employing offloaded collective operations. These capabilities can be used to reduce the effects of system noise, and when using nonblocking collective operations have the potential to hide the effects of application load imbalance.

99 GENERAL AND MISCELLANEOUS↗

A Fault-oblivious Extreme-scale Execution Environment

Exascale computing systems will provide a thousand-fold increase in parallelism and a proportional increase in failure rate relative to today's machines. Systems software for exascale machines must provide the infrastructure to support existing applications while simultaneously enabling efficient execution of new programming models that naturally express dynamic, adaptive, irregular computation; coupled simulations; and massive data analysis in a highly unreliable hardware environment with billions of threads of execution. We propose a new approach to the data and work distribution model provided by system software based on the unifying formalism of an abstract file system. The proposed hierarchical data model provides simple, familiar visibility and access to data structures through the file system hierarchy, while providing fault tolerance through selective redundancy. The hierarchical task model features work queues whose form and organization are represented as file system objects. Data and work are both first class entities. By exposing the relationships between data and work to the runtime system, information is available to optimize execution time and provide fault tolerance. The data distribution scheme provides replication (where desirable and possible) for fault tolerance and efficiency, and it is hierarchical to make it possible to take advantage of locality. The user, tools, and applications, including legacy applications, can interface with the data, work queues, and one another through the abstract file model. This runtime environment will provide multiple interfaces to support traditional Message Passing Interface applications, languages developed under DARPA's High Productivity Computing Systems program, as well as other, experimental programming models. We will validate our runtime system with pilot codes on existing platforms and will use simulation to validate for exascale-class platforms. In this final report, we summarize research results from the work done at the Ohio State University towards the larger goals of the project listed above.

97 MATHEMATICS AND COMPUTING↗

High-performance Computing Applied to Semantic Databases

To-date, the application of high-performance computing resources to Semantic Web data has largely focused on commodity hardware and distributed memory platforms. In this paper we make the case that more specialized hardware can offer superior scaling and close to an order of magnitude improvement in performance. In particular we examine the Cray XMT. Its key characteristics, a large, global shared-memory, and processors with a memory-latency tolerant design, offer an environment conducive to programming for the Semantic Web and have engendered results that far surpass current state of the art. We examine three fundamental pieces requisite for a fully functioning semantic database: dictionary encoding, RDFS inference, and query processing. We show scaling up to 512 processors (the largest configuration we had available), and the ability to process 20 billion triples completely in-memory.

99 GENERAL AND MISCELLANEOUS↗

Energy Conservation Through Cloned Execution Of Simulations

High-performance computing facilities used for scientific computing draw enormous energy, some of them consuming many megawatt-hours. Saving the energy consumption of computations on such facilities can dramatically reduce the total cost of their operation and help reduce environmental effects. Here, we focus on a way to reduce energy consumption in many ensembles of simulations. Using the method of simulation cloning to exploit parallelism while also significantly conserving the computational and memory requirements, we perform a detailed empirical study of energy consumed on a large supercomputer consisting of hardware accelerator cards (graphical processing units, GPUs). We build on previous insights from mathematical analysis and implementation of cloned simulations that result in computational and memory savings by several orders-of-magnitude. Using instrumentation to track the power drawn by thousands of accelerator cards, we report significant aggregate energy savings from cloned simulations.

Yoginath, Srikanth↗

Acceleration of Graph Neural Network-Based Prediction Models in Chemistry via Co-Design Optimization on Intelligence Processing Units

Atomic structure prediction and associated property calculations are the bedrock of chemical physics. Since high-fidelity ab initio modeling techniques for computing the structure and properties can be prohibitively expensive, this motivates the development of machine-learning (ML) models that make these predictions more efficiently. Training graph neural networks over large atomistic databases introduces unique computational challenges such as the need to process millions of small graphs with variable size and support communication patterns that are distinct from learning over large graphs such as social networks. We demonstrate a novel hardware-software co-design approach to scale up the training of atomistic graph neural networks (GNN) for structure and property prediction. First, to eliminate redundant computation and memory associated with alternative padding techniques and to improve throughput via minimizing communication, we formulate the effective coalescing of the batches of variable-size atomistic graphs as the bin packing problem and introduce a hardware-agnostic algorithm to pack these batches. In addition, we propose hardware-specific optimizations including a planner and vectorization for the gather-scatter operations targeted for Graphcore’s Intelligence Processing Unit (IPU), as well as model-specific optimizations such as merged communication collectives and optimized softplus. Putting these all together, we demonstrate the effectiveness of the proposed co-design approach by providing an implementation of a well-established atomistic GNN on the Graphcore IPUs. We evaluate the training performance on multiple atomistic graph databases with varying degrees of graph counts, sizes and sparsity. Here, we demonstrate that such a co-design approach can reduce the training time of atomistic GNNs and can improve the performance by up to 1.5× compared to the baseline implementation of the model on the IPUs. Additionally, we compare our IPU implementation with a Nvidia GPU-based implementation and show that our atomistic GNN implementation on the IPUs can run 1.8× faster on average compared to the execution time on the GPUs.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Understanding Line Losses and Transformer Losses in Rural Isolated Distribution Systems

Rural, isolated power systems in the mainland U.S. and in states like Alaska and Hawaii are powered by assets like diesel generators. These rural, isolated power systems also cannot operate at the higher band of medium voltage (like 69kV). They are primarily in the 12 to 14 kV range to keep the cost of the distribution investments lower. Because of this mid-band medium voltage range, the line losses and distribution transformers losses consume significant diesel consumption (almost 10 percent of the peak load). This work considers one such power system powering an isolated system and presents key findings online losses, and transformer losses. Understanding and documenting the impacts is critical for these communities operating their power systems and take actions to reduce expensive diesel consumption. In this paper, we will present one such typical grid and model it in electromagnetic transients (EMT) domain. We used the tower structure, under ground cabling installation to develop high fidelity models of lines. We also used high fidelity models of distribution transformers to present the no-load losses and full load loses. We will also present technical solutions available commercially off-the-shelf to reduce these losses and reduce diesel consumption. This work will be a primer for communities to understand the technical challenges and to understand the possible solution available to solve such challenges for rural, isolated power system operators.

blackstart↗

Implementing machine learning methods on QICK hardware for qubit readout & control

Quantum readout and control is a fundamental aspect of quantum computing that requires accurate measurement of qubit states. Errors emerge in all stages, from initialization to readout, and identifying errors in post-processing necessitates resource-intensive statistical analysis. In our work, we use a lightweight fully-connected neural network (NN) to classify states of a transmon system with no prior processing. Our NN accelerator yields higher fidelities (92%) than the classical matched filter method (84%). By exploiting the natural parallelism of NNs and their placement near the source of data on field-programmable gate arrays (FPGAs), we can achieve ultra-low latency on the Quantum Instrumentation Control Kit (QICK). Integrating machine learning methods on QICK opens several pathways for efficient real-time processing of quantum circuits.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Computational Tools for Additive Manufacture of Tailored Microstructure and Properties

Additive manufacturing has the potential to revolutionize industrial hardware and unlock efficiency gains through the fabrication of geometries and architectures not possible by conventional processing. Currently most additive builds use a single set of process parameters (e.g. laser power and scan speed) which results in a part with a homogenous microstructure that provides a singular performance level. To move beyond this state, Raytheon Technologies Research Center worked to create a set of computational tools to track material evolution through each step of the additive process. Computational fluid dynamics and phase field models for microstructure evolution as a function of processing parameters, and crystal plasticity models fully coupling microstructure and mechanical properties for performance predictions were leveraged to establish a connection between additive parameters and the final microstructure. This framework was utilized to tailor spatially-varying mechanical properties in a part by appropriately controlling the microstructure evolution during the additive process. Specifically, a turbine blade was 3D printed from nickel superalloy IN718 using laser powder bed fusion with coarse grains in the airfoil section which experiences the highest temperatures and is creep limited while finer grains were printed in the root of the blade which experience higher stresses but at lower temperatures and is therefore fatigue limited. The benefit of being able to intentionally insert coarse grains in the high temperature region of the blade was showcased with a microstructure sensitive creep model that indicates longer creep life for coarser grains.

20 FOSSIL-FUELED POWER PLANTS↗

FOX: A Fault-Oblivious Extreme-Scale Execution Environment Boston University Final Report Project Number: DE-SC0005365

Exascale computing systems will provide a thousand-fold increase in parallelism and a proportional increase in failure rate relative to today's machines. Systems software for exascale machines must provide the infrastructure to support existing applications while simultaneously enabling efficient execution of new programming models that naturally express dynamic, adaptive, irregular computation; coupled simulations; and massive data analysis in a highly unreliable hardware environment with billions of threads of execution. The FOX project explored systems software and runtime support for a new approach to the data and work distribution for fault oblivious application execution. Our major OS work at Boston University focused on developing a new light-weight operating systems model that provides an appropriate context for both multi-core and multi-node application development. This work is discussed in section 1. Early on in the FOX project BU developed infrastructure for prototyping dynamic HPC environments in which the sets of nodes that an application is run on can be dynamically grown or shrunk. This work was an extension of the Kittyhawk project and is discussed in section 2. Section 3 documents the publications and software repositories that we have produced. To put our work in context of the complete FOX project contribution we include in section 4 an extended version of a paper that documents the complete work of the FOX team.

97 MATHEMATICS AND COMPUTING↗

An Integrated Development Environment for Adiabatic Quantum Programming

Adiabatic quantum computing is a promising route to the computational power afforded by quantum information processing. The recent availability of adiabatic hardware raises the question of how well quantum programs perform. Benchmarking behavior is challenging since the multiple steps to synthesize an adiabatic quantum program are highly tunable. We present an adiabatic quantum programming environment called JADE that provides control over all the steps taken during program development. JADE captures the workflow needed to rigorously benchmark performance while also allowing a variety of problem types, programming techniques, and processor configurations. We have also integrated JADE with a quantum simulation engine that enables program profiling using numerical calculation. The computational engine supports plug-ins for simulation methodologies tailored to various metrics and computing resources. We present the design, integration, and deployment of JADE and discuss its use for benchmarking adiabatic quantum programs.

ISING MODEL↗

SmartApps: Middle-ware for Adaptive Applications on Reconfigurable Platforms

There are several reasons why the performance of current distributed and heterogeneous systems is often disappointing. For example, the characteristics of the application may be input sensitive and evolve during execution causing dramatic changes in memory reference patterns, resource requirements, or degree of concurrency between different phases of the computation. Or, the system may change dynamically with nodes failing or appearing, some network links severed and other links established with different latencies and bandwidths. Another important reason for poor performance is the fairly compartmentalized approach to optimization: applications, compilers, operating systems and hardware configurations are designed and optimized in isolation and without the knowledge of instance specific information and needs of a running application. There is too little information flow across these boundaries and no global optimization is even attempted. For example, most operating systems services like paging, virtual-to-physical page mapping, I/O, or data layout in disks, provide little or no application customization. Similarly, the off-the-shelf hardware used by most commercial systems is optimized to give best average-case performance. To address this problem, we have proposed application-centric computing, or Smart Applications (SAS). In the SAS executable, the compiler embeds most run-time system services, and a performance-optimizing feedback loop that monitors the application's performance and adaptively reconfigures the application and the OS/system platform. At run-time, after incorporating the code's input and determining the system's resources and state, the SAS performs an *instance* specific optimization, which is more tractable than a global generic optimization between application, OS and system. The overriding philosophy of SAS is ``measure, compare, and adapt if beneficial.'' That is, the application will continually monitor its performance and the available resources to determine if, and by how much, performance could be improved if the application was restructured. Then, if the potential performance benefit outweighs the projected overhead costs, the application will restructure itself and the underlying system accordingly. The SAS framework includes performance monitoring and modeling components and mechanisms for performing the actual restructuring at various levels including: (i) algorithmic adaptation, (ii) run-time software optimization (e.g., input sensitivity analysis, etc.), (iii) tuning reconfigurable OS services (scheduling policy, page size, etc), and (iv) system configuration (e.g., selecting which, and how many, computational resources to use). SmartApps is being developed in the STAPL infrastructure. STAPL (the Standard Template Adaptive Parallel Library) is a framework for developing highly-optimizable, adaptable, and portable parallel and distributed applications. It consists of a relatively new and still evolving collection of generic parallel algorithms and distributed containers and a run-time system (RTS) through which the application and compiler interact with the OS and hardware.

97 MATHEMATICS AND COMPUTING↗

Frequency Response Analysis to Monitor and Identify Changes in the Impedance of a Photovoltaic Panel Measured Online using a Power Optimizer

Photovoltaic (PV) cells are generally modeled as a current source due to photocurrent, p-n junction diodes with parasitic resistance, capacitance, and inductance. This paper proposes online frequency response analysis (FRA) to measure the impedance of a PV panel using an existing panel-level power optimizer in a PV system. The algorithm will actively perturb a small signal into a 300 W rooftop PV panel and compute its small signal impedance. The technology discussed is easy to incorporate, requires no additional hardware, doesn't alter the stability of the system, and is implemented at a steady-state point. The power optimizer initially stabilizes at an operating point and then perturbs the PV current via FRA and computes PV panel impedance. The relative standard deviation test conducted indoors under 300 W/m 2 illumination on a PV panel shows a less than 5% error rate in PV panel impedance magnitude and phase is measured using a power optimizer.

14 SOLAR ENERGY↗

Modularity and adaptability

The primary question I am addressing is how to construct simulation tools, in this case pri marily PDE simulations, that can be made to efficiently use a future computer whose construction and parameters are not known exactly, but whose general construction can be inferred from current prevailing hardware trends. These trends inform our best estimates of the likely composition of such a next generation machine, and the outlook is that the computers of the future will be highly het erogenous, will penalize excessive communication (particularly with far away pieces of the machine), can likely be tuned to trade pure speed for reliability, and may experience frequent disruptions. In such an environment, my position as to what mathematical research should be performed to aid in this endeavor can be summed up as a need to stress the following points: • Increased modularity of the simulation tool, • Retention of original problem definition as much as possible, and • Provide and analyze flexible alternatives.

97 MATHEMATICS AND COMPUTING↗

Searching for Physics Beyond the Standard Model: Strongly-Coupled Field Theories at the Intensity and Energy Frontiers

This proposal is to develop the software and algorithmic infrastructure needed for the numerical study of quantum chromodynamics (QCD), and of theories that have been proposed to describe physics beyond the Standard Model (BSM) of high energy physics, on current and future computers. This infrastructure will enable users (1) to improve the accuracy of QCD calculations to the point where they no longer limit what can be learned from high-precision experiments that seek to test the Standard Model, and (2) to determine the predictions of BSM theories in order to understand which of them are consistent with the data that will soon be available from the LHC. Work will include the extension and optimizations of community codes for the next generation of leadership class computers, the IBM Blue Gene/Q and the Cray XE/XK, and for the dedicated hardware funded for our field by the Department of Energy. Members of our collaboration at Brookhaven National Laboratory and Columbia University worked on the design of the Blue Gene/Q, and have begun to develop software for it. Under this grant we will build upon their experience to produce high-efficiency production codes for this machine. Cray XE/XK computers with many thousands of GPU accelerators will soon be available, and the dedicated commodity clusters we obtain with DOE funding include growing numbers of GPUs. We will work with our partners in NVIDIA's Emerging Technology group to scale our existing software to thousands of GPUs, and to produce highly efficient production codes for these machines. Work under this grant will also include the development of new algorithms for the effective use of heterogeneous computers, and their integration into our codes. It will include improvements of Krylov solvers and the development of new multigrid methods in collaboration with members of the FASTMath SciDAC Institute, using their HYPRE framework, as well as work on improved symplectic integrators.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A Report on Simulation-Driven Reliability and Failure Analysis of Large-Scale Storage Systems

High-performance computing (HPC) storage systems provide data availability and reliability using various hardware and software fault tolerance techniques. Usually, reliability and availability are calculated at the subsystem or component level using limited metrics such as, mean time to failure (MTTF) or mean time to data loss (MTTDL). This often means settling on simple and disconnected failure models (such as exponential failure rate) to achieve tractable and close-formed solutions. However, such models have been shown to be insufficient in assessing end-to-end storage system reliability and availability. We propose a generic simulation framework aimed at analyzing the reliability and availability of storage systems at scale, and investigating what-if scenarios. The framework is designed for an end-to-end storage system, accommodating the various components and subsystems, their interconnections, failure patterns and propagation, and performs dependency analysis to capture a wide-range of failure cases. We evaluate the framework against a large-scale storage system that is in production and analyze its failure projections toward and beyond the end of lifecycle. We also examine the potential operational impact by studying how different types of components affect the overall system reliability and availability, and present the preliminary results

97 MATHEMATICS AND COMPUTING↗

Hardware-Efficient Quantum Random Access Memory with Hybrid Quantum Acoustic Systems

Hybrid quantum systems in which acoustic resonators couple to superconducting qubits are promising quantum information platforms. High quality factors and small mode volumes make acoustic modes ideal quantum memories, while the qubit-phonon coupling enables the initialization and manipulation of quantum states. In this paper, we present a scheme for quantum computing with multimode quantum acoustic systems, and based on this scheme, propose a hardware-efficient implementation of a quantum random access memory (QRAM). Quantum information is stored in high-$\textit{Q}$ phonon modes, and couplings between modes are engineered by applying off-resonant drives to a transmon qubit. In comparison to existing proposals that involve directly exciting the qubit, this scheme can offer a substantial improvement in gate fidelity for long-lived acoustic modes. We show how these engineered phonon-phonon couplings can be used to access data in superposition according to the state of designated address modes—implementing a QRAM on a single chip.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Scalable FPGA Accelerator for Deep Convolutional Neural Networks with Stochastic Streaming

Here, FPGA-based heterogeneous computing platform, due to its extreme logic reconfigurability, emerges to be a strong contender as computing fabric in modern AI. As a result, various FPGA-based accelerators for deep CNN—the key driver of modern AI—have been proposed due to their advantages of high performance, reconfigurability, and fast development round, etc. In general, the consensus among researchers is that, although FPGA-based accelerator can achieve much higher energy efficiency, its raw computing performance lags behind when compared with GPUs with similar logic density. In this paper, we develop an alternative methodology to efficiently implement CNNs with FPGAs that outperform GPUs in terms of both power consumption and performance. Our key idea is to design a scalable hardware architecture and circuit design for large-scale CNNs that leverages a stochastic-based computing principle. Specifically, there are three major performance advantages. First, all key components of our deep learning CNN are designed and implemented to compute stochastically, thus achieving excellent computing performance and energy efficiency. Second, because our proposed CNN architecture enables a stream-mode computing, all of its stages can process even the partial results from preceding stages, therefore not incurring unnecessary latency due to data dependency. Finally, our FPGA-based deep CNN also provides a superior hardware scalability when compared with conventional FPGA implementations by reducing the bandwidth requirement between layers. The results show that our proposed CNN architecture significantly outperforms all previous FPGA-based deep CNN implementation approaches. It achieves 1.58x more GOPS, 6.42x more GOPS/Slice, and 10.92x more GOPS/W when compared with state-of-the-art CNN architecture. The top-5 accuracy of stochastic VGG-16 CNN is 86.77 percent with 18.91 fps frame rate.

97 MATHEMATICS AND COMPUTING↗

ARENA: Asynchronous Reconfigurable Accelerator Ring to Enable Data-Centric Parallel Computing

The next generation HPC and data centers are likely to be reconfigurable and data-centric due to the trend of hardware specialization and the emergence of data-driven applications. In this work, we propose ARENA – an asynchronous reconfigurable accelerator ring architecture as a potential scenario on how the future HPC and data centers will be like. Despite using the coarse-grained reconfigurable arrays (CGRAs) as the substrate platform, our key contribution is not only the CGRA-cluster design itself, but also the ensemble of a new architecture and programming model that enables asynchronous tasking across a cluster of reconfigurable nodes, so as to bring specialized computation to the data rather than the reverse. We presume distributed data storage without asserting any prior knowledge on the data distribution. Hardware specialization occurs at runtime when a task finds the majority of data it requires are available at the present node. In other words, we dynamically generate specialized CGRA accelerators where the data reside. The asynchronous tasking for bringing computation to data is achieved by circulating the task token, which describes the dataflow graphs to be executed for a task, among the CGRA cluster connected by a fast ring network. Evaluations on a set of HPC and data-driven applications across different domains show that ARENA can provide better parallel scalability with reduced data movement (53.9 percent). Compared with contemporary compute-centric parallel models, ARENA can bring on average 4.37× speedup. The synthesized CGRAs and their task-dispatchers only occupy 2.93mm 2 chip area under 45nm process technology and can run at 800MHz with on average 759.8mW power consumption. ARENA also supports the concurrent execution of multi-applications, offering ideal architectural support for future high-performance parallel computing and data analytics systems.

97 MATHEMATICS AND COMPUTING↗