Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “interconnection networks”

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 181 records · Page 10

A Quantitative Study of Deep Learning Training on Heterogeneous Supercomputers

Deep learning (DL) has become a key technique for solving complex problems in scientific research and discovery. DL training for science is substantially challenging because it has to deal with massive quantities of multi-dimensional data. High- performance computing (HPC) supercomputers are increasingly being employed for meeting the exponentially growing demand for DL. Multiple GPUs and high-speed interconnect network are needed for supporting DL on HPC systems. However, the excessive use of GPUs without considering effective benefits leads to inefficient resource utilization of these expensive setups. In this paper, we conduct a quantitative analysis to gauge the efficacy of DL workloads on the latest HPC system and identify viability of next-generation DL-optimized heterogeneous supercomputers for enabling researchers to develop more efficient resource management and distributed DL middleware. We evaluate well- known DL models with large-scale datasets using the popular TensorFlow framework, and provide a thorough evaluation in- cluding scalability, accuracy, variability, storage resource, GPU- GPU/GPU-CPU data transfer, and GPU utilization. Our analysis reveals that the latest heterogeneous supercomputing cluster shows varying performance trend as compared to the existing literature for single- and multi-node training. To the best of our knowledge, this is the first work to conduct such a quantitative and comprehensive study of DL training on a supercomputing system with multiple GPUs.

Han, Jingoo↗

Correlation between Generator Trips and Locational Marginal Prices (LMPs)

Generally, disturbances in an interconnected network have an influence the stability and the economy of the power grid. Some phenomenon such as load shedding have proven to have a huge impact on real time electricity prices. Generation trip is considered one of the most common events in a power grid that may change the wholesale electricity market. In this paper, a locational marginal price (LMP) measure is used to notice how the electricity prices change with generator trip events at different locations of U.S. power grid. The study results help to find whether there is a clear correlation between generator trips and the changes in the electricity real-time market.

Alharbi, Abdulelah↗

A Hierarchical Multi-Agent Based Anomaly Detection for Wide-Area Protection in Smart Grid

Future smart grid capabilities provide assurance to expand the advanced information and communication technologies to evolve into densely interconnected cyber physical system. Remedial Action Scheme (RAS), widely used for wide-area protection, relies on the interconnected networks and data sharing devices, which are exposed to the multitude of vulnerabilities. This paper presents our proposed approach to developing multi-agent based RAS scheme against the system-aware stealthy cyber-attacks. Specifically, we propose the two-level hierarchical architecture which consists of distributed local RAS controllers (RAScs) as local agents, operating at different zones/ areas, which are constantly monitored by an overseer, the central agent. The local controllers receive local and randomly changing outside zonal measurements and cyclically forwards to the overseer. The overseer identifies the corrupted controller using the anomaly detection algorithm which processes the measurements coming from the local controllers, compute measurement errors using local and outside zonal measurements, perform validation checks, and finally detect anomalies based on the two-step verification. Next, as a proof of concept, we have implemented and validated the proposed methodology in cyber physical environment at Iowa State’s PowerCyber testbed. We have also implemented the coordinated attack vectors which involve corrupting the local controller and later performing stealthy attacks on the system’s generator. We have evaluated its performance during the online testing in terms of detection rate and Iatency. The experimental results show that it is efficient in detecting different classes of attacks, including ramp and pulse attacks.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Uncovering Droop Control Laws Embedded Within the Nonlinear Dynamics of Van der Pol Oscillators

This paper examines the dynamics of power-electronic inverters in islanded microgrids that are controlled to emulate the dynamics of Van der Pol oscillators. The general strategy of controlling inverters to emulate the behavior of nonlinear oscillators presents a compelling time-domain alternative to ubiquitous droop control methods which presume the existence of a quasistationary sinusoidal steady state and operate on phasor quantities. We present two main results in this paper. First, by leveraging the method of periodic averaging, we demonstrate that droop laws are intrinsically embedded within a slower time scale in the nonlinear dynamics of Van der Pol oscillators. Second, we establish the global convergence of amplitude and phase dynamics in a resistive network interconnecting inverters controlled as Van der Pol oscillators. Furthermore, under a set of nonrestrictive decoupling approximations, we derive sufficient conditions for local exponential stability of desirable equilibria of the linearized amplitude and phase dynamics.

24 POWER TRANSMISSION AND DISTRIBUTION↗

A dynamic protein interactome drives energy conservation and electron flux in Thermococcus kodakarensis

ABSTRACT Life is supported by energy gains fueled by catabolism of a wide range of substrates, each reliant on the selective partitioning of electrons through redox ( red uction and ox idation) reactions. Electron flux through tunable and regulated protein interactions provides dynamic routes for energy conservation, but how electron flux is regulated in vivo , particularly for archaeal metabolisms that support rapid growth at the thermodynamic limits of life, is poorly understood. Identification of bona fide in vivo protein assemblies and how such assemblies dictate the totality of electron flux is critical to our understanding of the regulation imposed on metabolism, energy production, and energy conservation. Here, 25 key proteins in central metabolic redox pathways in the model, genetically accessible, hyperthermophilic archaeon Thermococcus kodakarensis , were purified to reveal an extensive, dynamic, and tightly interconnected network of protein interactions that responds to environmental cues (such as the availability of various reductive sinks) to direct electron flux to maximize energetic gains. Interactions connecting disparate functions suggest many catabolic and anabolic activities occur in spatial proximity in vivo , and while protein complexes have been historically defined under optimal conditions, many of these complexes appear to maintain alternative partnerships in changing conditions. The totality of the results obtained redefines our understanding of in vivo assemblies driving ancient metabolic strategies supporting the growth of modern Archaea. IMPORTANCE Given the potential for rational genetic manipulations of biofuel- and biotech-promising archaea to yield transformative results for major markets, it is a priority to define how the metabolisms of such species are controlled, at least in part, by in vivo protein assemblies, and from such, define routes of energy flux that can be most efficiently altered toward biofuel or biotechnological gains. Proteinaceous electron carriers (PECs, such as ferredoxins) offer the potential for specific protein–protein interactions to coordinate selective reductive flow. Employing the model, genetically accessible, hyperthermophilic archaeon, Thermococcus kodakarensis , we establish the metabolic protein interactome of 25 key redox proteins, revealing that each redox active protein has a dynamic partnership profile, suggesting catabolic and anabolic activities may occur in concert and in temporal and spatial proximity in vivo . These results reveal critical importance in evaluating the newly identified partnerships and their role and utility in providing regulated redox flux in T. kodakarensis .

Williams, Sere A. (ORCID:0000000235509590)↗

Scalable Asynchronous Domain Decomposition Solvers

We discuss how parallel implementations of linear iterative solvers generally alternate between phases of data exchange and phases of local computation. Increasingly large problem sizes and more heterogeneous compute architectures make load balancing and the design of low latency network interconnects that are able to satisfy the communication requirements of linear solvers very challenging tasks. In particular, global communication patterns such as inner products become increasingly limiting at scale. We explore the use of asynchronous communication based on one-sided Message Passing Interface primitives in the context of domain decomposition solvers. In particular, a scalable asynchronous two-level Schwarz method is presented. We discuss practical issues encountered in the development of a scalable solver and show experimental results obtained on a state-of-the-art supercomputer system that illustrate the benefits of asynchronous solvers in load balanced as well as load imbalanced scenarios. Using the novel method, we can observe speedups of up to four times over its classical synchronous equivalent.

97 MATHEMATICS AND COMPUTING↗

On the Strong Scaling of the Spectral Element Solver Nek5000 on Petascale Systems. In: EASC '16 Proceedings of the Exascale Applications and Software Conference 2016, Article No. 5

The present work is targeted at performing a strong scaling study of the high-order spectral element fluid dynamics solver Nek5000. Prior studies such as [5] indicated a recommendable metric for strong scalability from a theoretical viewpoint, which we test here extensively on three parallel machines with different performance characteristics and interconnect networks, namely Mira (IBM Blue Gene/Q), Beskow (Cray XC40) and Titan (Cray XK7). The test cases considered for the simulations correspond to a turbulent flow in a straight pipe at four different friction Reynolds numbers Reτ = 180, 360, 550 and 1000. Considering the linear model for parallel communication we quantify the machine characteristics in order to better assess the scaling behaviors of the code. Subsequently sampling and profiling tools are used to measure the computation and communication times over a large range of compute cores. We also study the effect of the two coarse grid solvers XXT and AMG on the computational time. Super-linear scaling due to a reduction in cache misses is observed on each computer. The strong scaling limit is attained for roughly 5000-10,000 degrees of freedom per core on Mira, 30,000 - 50,0000 on Beskow, with only a small impact of the problem size for both machines, and ranges between 10,000 and 220,000 depending on the problem size on Titan. This work aims at being a reference for Nek5000 users and also serves as a basis for potential issues to address as the community heads towards exascale supercomputers.

Offermans, Nicolas↗

Thread Batching for High-performance Energy-efficient GPU Memory Design

Massive multi-threading in GPU imposes tremendous pressure on memory subsystems. Due to rapid growth in thread-level parallelism of GPU and slowly improved peak memory bandwidth, memory becomes a bottleneck of GPU’s performance and energy efficiency. In this article, we propose an integrated architectural scheme to optimize the memory accesses and therefore boost the performance and energy efficiency of GPU. First, we propose a thread batch enabled memory partitioning (TEMP) to improve GPU memory access parallelism. In particular, TEMP groups multiple thread blocks that share the same set of pages into a thread batch and applies a page coloring mechanism to bound each stream multiprocessor (SM) to the dedicated memory banks. After that, TEMP dispatches the thread batch to an SM to ensure high-parallel memory-access streaming from the different thread blocks. Second, a thread batch-aware scheduling (TBAS) scheme is introduced to improve the GPU memory access locality and to reduce the contention on memory controllers and interconnection networks. Experimental results show that the integration of TEMP and TBAS can achieve up to 10.3% performance improvement and 11.3% DRAM energy reduction across diverse GPU applications. We also evaluate the performance interference of the mixed CPU+GPU workloads when they are run on a heterogeneous system that employs our proposed schemes. Our results show that a simple solution can effectively ensure the efficient execution of both GPU and CPU applications.

Computer Science↗

Dragonview

Dragonview is an interactive visual analytics tool to visualize the nodes and links of a supercomputer that uses the dragonfly topology for its interconnection network. It can be used to map different data such as number of packets or bytes or stalls on the network links and job IDs or other on-node metrics on the router or node.

Source record↗

Design and implementation of I/O performance prediction scheme on HPC systems through large-scale log analysis

Abstract Large-scale high performance computing (HPC) systems typically consist of many thousands of CPUs and storage units used by hundreds to thousands of users simultaneously. Applications from large numbers of users have diverse characteristics, such as varying computation, communication, memory, and I/O intensity. A good understanding of the performance characteristics of each user application is important for job scheduling and resource provisioning. Among these performance characteristics, I/O performance is becoming increasingly important as data sizes rapidly increase and large-scale applications, such as simulation and model training, are widely adopted. However, predicting I/O performance is difficult because I/O systems are shared among all users and involve many layers of software and hardware stack, including the application, network interconnect, operating system, file system, and storage devices. Furthermore, updates to these layers and changes in system management policy can significantly alter the I/O behavior of applications and the entire system. To improve the prediction of the I/O performance on HPC systems, we propose integrating information from several different system logs and developing a regression-based approach to predict the I/O performance. Our proposed scheme can dynamically select the most relevant features from the log entries using various feature selection algorithms and scoring functions, and can automatically select the regression algorithm with the best accuracy for the prediction task. The evaluation results show that our proposed scheme can predict the write performance with up to 90% prediction accuracy and the read performance with up to 99% prediction accuracy using the real logs from the Cori supercomputer system at NERSC.

97 MATHEMATICS AND COMPUTING↗

Integration of energy storage with diesel generation in remote communities

Highlights Battery energy storage may improve energy efficiency and reliability of hybrid energy systems composed by diesel and solar photovoltaic power generators serving isolated communities. In projects aiming update of power plants serving electrically isolated communities with redundant diesel generation, battery energy storage can improve overall economic performance of power supply system by reducing fuel usage, decreasing capital costs by replacing redundant diesel generation units, and increasing generator system life by shortening yearly runtime. Fast-acting battery energy storage systems with grid-forming inverters might have potential for improving drastically the reliability indices of isolated communities currently supplied by diesel generation. Abstract This paper will highlight unique challenges and opportunities with regard to energy storage utilization in remote, self-sustaining communities. The energy management of such areas has unique concerns. Diesel generation is often the go-to power source in these scenarios, but these systems are not devoid of issues. Without dedicated maintenance crews as in large, interconnected network areas, minor interruptions can be frequent and invasive not only for those who lose power, but also for those in the community that must then correct any faults. Although the immediate financial benefits are perhaps not readily apparent, energy storage could be used to address concerns related to reliability, automation, fuel supply concerns, generator degradation, solar utilization, and, yes, fuel costs to name a few. These ideas are shown through a case study of the Levelock Village of Alaska. Currently, the community is faced with high diesel prices and a difficult supply chain, which makes temporary loss of power very common and reductions in fuel consumption very impactful. This study will investigate the benefits that an energy storage system could bring to the overall system life, fuel costs, and reliability of the power supply. The variable efficiency of the generators, impact of startup/shutdown process, and low-load operation concerns are considered. The technological benefits of the combined system will be explored for various scenarios of future diesel prices and technology maintenance/replacement costs as well as for the avoidance of power interruptions that are so common in the community currently. Graphic abstract Discussion In several cases, energy storage can provide a means to promote energy equity by improving remote communities’ power supply reliability to levels closer to what the average urban consumer experiences at a reduced cost compared to transmission buildout. Furthermore, energy equity represents a hard-to-quantify benefit achieved by the integration of energy storage to isolated power systems of under-served communities, which suggests that the financial aspects of such projects should be questioned as the main performance criterion. To improve battery energy storage system valuation for diesel-based power systems, integration analysis must be holistic and go beyond fuel savings to capture every value stream possible.

25 ENERGY STORAGE↗

Solution-Processable Transparent Conductive Hole Injection Electrode for OLED SSL

An interconnected network of silver nanowires has been used as transparent anode in OLED devices. This layer was deposited by spin-coating and slot-die coating from an aqueous nanowire suspension. The sheet resistance of the film was 10ohms/sq with a transmission (including the glass substrate) of higher than 85%. The first phase of the project focused on the implementation of this nanowire layer with a hole-injection-layer (HIL) which has been developed at Plextronics and has been shown to provide good stability and efficiency in conventional OLED devices. We modified the HIL solution such that it coated reasonably well with suitable surface morphology so that actual devices can be manufactured. During the second phase we investigated the hole-injection and stability of hole-onlydevices. We determined that the use of the nanowire network as anode does not introduce an additional degradation mechanism since the observed device characteristics did not differ from those made with ITO anode. We then proceeded to make actual OLED devices with this nanowire / HIL stack and achieved device characteristics similar state-of-the-art OLED devices with a single junction. In order to gain traction with potential OLED manufacturers, we decided to contract Novaled to prepare large-area demonstrators for us. For these devices, we used an allevaporated stack, i.e. we did use Novaled's HIL material instead of Plextronics. We successfully fabricated demonstrators with an area of 25cm2 with a double or triple junction stack. Minor stack optimizations were necessary to achieve efficacies and lifetime equivalent with ITO devices made with the same devices stack. Due to the reduced microcavity effect, the color of the emitted light is significantly more stable with respect to the viewing angle compared to ITO devices. This fact in conjunction with the promise of lower production cost due to the elimination of the ITO sputtering process and the direct patterning of the anode layer are the obvious advantages of this technology. The project has shown that this nanowire technology is a viable option to achieve OLED devices with good lifetime and efficiency and we are currently working with manufacturers to utilize this technology in a production setting.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

A Framework for Adaptable Operating and Runtime Systems

The emergence of new classes of HPC systems where performance improvement is enabled by Moore’s Law for technology is manifest through multi-core-based architectures including specialized GPU structures. Operating systems were originally designed for control of uniprocessor systems. By the 1980s multiprogramming, virtual memory, and network interconnection were integral services incorporated as part of most modern computers. HPC operating systems were primarily derivatives of the Unix model with Linux dominating the Top-500 list. The use of Linux for commodity clusters was first pioneered by the NASA Beowulf Project. However, the rapid increase in number of cores to achieve performance gain through technology advances has exposed the limitations of POSIX general-purpose operating systems in scaling and efficiency. This project was undertaken through the leadership of Sandia National Laboratories and in partnership of the University of New Mexico to investigate the alternative of composable lightweight kernels on scalable HPC architectures to achieve superior performance for a wide range of applications. The use of composable operating systems is intended to provide a minimalist set of services specifically required by a given application to preclude overheads and operational uncertainties (“OS noise”) that have been demonstrated to degrade efficiency and operational consistency. This project was undertaken as an exploration to investigate possible strategies and methods for composable lightweight kernel operating systems towards support for extreme scale systems.

97 MATHEMATICS AND COMPUTING↗

Automatic Energy Schemes for High Performance Applications

Although high-performance computing traditionally focuses on the efficient execution of large-scale applications, both energy and power have become critical concerns when approaching exascale. Drastic increases in the power consumption of supercomputers affect significantly their operating costs and failure rates. In modern microprocessor architectures, equipped with dynamic voltage and frequency scaling (DVFS) and CPU clock modulation (throttling), the power consumption may be controlled in software. Additionally, network interconnect, such as Infiniband, may be exploited to maximize energy savings while the application performance loss and frequency switching overheads must be carefully balanced. This work first studies two important collective communication operations, all-to-all and allgather and proposes energy saving strategies on the per-call basis. Next, it targets point-to-point communications to group them into phases and apply frequency scaling to them to save energy by exploiting the architectural and communication stalls. Finally, it proposes an automatic runtime system which combines both collective and point-to-point communications into phases, and applies throttling to them apart from DVFS to maximize energy savings. The experimental results are presented for NAS parallel benchmark problems as well as for the realistic parallel electronic structure calculations performed by the widely used quantum chemistry package GAMESS. Close to the maximum energy savings were obtained with a substantially low performance loss on the given platform.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Cyber Security Analysis for Nuclear Reactor Control Systems (Final Technical Report)

This project investigated the cyber-security impacts of moving from an all analog, point-to-point, instrumentation and control (I&C) system to a digital I&C system based on Modbus and a shared communication medium. A formalism called a hybrid attack graph was expanded to support the nuclear research reactor system. The hybrid attack graph allows one to check a system for vulnerabilities, in this case cyber-security vulnerabilities, and to document the attack vectors (scenarios) causing those vulnerabilities. In parallel, a simulation of the system was developed to model both the physical reactor parameters and operations, as well as the network interconnects and communications. This simulation platform was modeled on the nuclear research reactor located at Washington State University. The simulation platform provided a sandbox to evaluate and quantify the impact of identified and proposed vulnerabilities in the system and to determine the effectiveness of countermeasures at stopping these attacks. The simulation and hybrid attack graph tools were integrated to provide a streamlined process of generating attack scenarios, playing those scenarios out in the simulation, and then analyzing the results to correlate system state to states in the hybrid attack graph. This process was used to (1) quantify the impact of attack scenarios and (2) to determine if the system moved through the hybrid attack graph as anticipated. The hybrid attack graph tool was extended and customized to produce a tool to automatically identify critical assets (CAs) and critical digital assets (CDAs) as defined by NRC Regulatory Guide 5.71. This tool was verified using the nuclear research reactor at Washington State University. Finally, a series of educational modules covering the findings of the different aspects of this research have been created.

97 MATHEMATICS AND COMPUTING↗

Digitalization in Electric Power Systems and Regulation: A Primer

This report defines grid digitalization, provides concrete examples, and assesses five elements of digital systems that have evolved over the past three decades relevant to the electric distribution system: (1) network interconnection, (2) interoperability, (3) modularity, (4) open source, and (5) automation. This report also discusses the benefits and challenges of digitalization and provides a set of actions regulators can take to support the transition and realize the benefits of the transition.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Tailoring Carbide Dispersed Steels: A Path to Increased Strength and Hydrogen Tolerance

The use of transition metal carbides is reported for use as a hydrogen trapping mechanism for ferritic and austenitic steel materials. The program combined computational modeling and simulations to guide experiments towards candidate metal carbide traps, both for interfacial and interior trapping. It was found that interfacial trapping is less effective than interior trapping, with the group IVB transition metal carbides being the most effect internal traps with a loss of carbon. The sub-stoichiometric rocksalt structure accommodate the hydrogen atoms in its octahedral interstices. Using percolation theory, carbon loss of approximately 25% or more was sufficient to ensure an interconnected network of vacancies for such trapping from the surface to the internal sites within the carbide. Using this as a guide, the program developed a means to provide a uniform dispersion of ZrC nanoparticles with either Fe or 304L micron-scale powders which was then consolidated by direct current sintering. Electrolytic hydrogen diffusivity studies confirmed the reduction of hydrogen diffusion in the matrix with increasing ZrC content, which was a linear response over the sample range studied (0.01 to 1.0 wt.%). The consolidated material was micro-tensile tested in either a non-hydrogen or hydrogen charge condition and compared to a control with no carbides. Additions up to 0.05 wt.% ZrC increased the yield strength with no loss in ductility in either the non-hydrogen or hydrogen tested condition. ZrC concentrations above this amount further increased the yield strength at the expense of ductility. While these samples had a lower absolute ductility value prior to failure, the relative change in ductility between the non-hydrogen and hydrogen charge states was less for the carbides than that of the control. Metal-rich ZrC nanoparticles were fabricated through a conformal coating process yielding ZrC0.66 particles that were then incorporated into a metal matrix. Notch fatigue testing in a hydrogen environment was conducted where the number of cycles to failure was found to be less in the control than that of the carbide addition. However, the spread in experimental data and the number of samples tested limits a conclusive outcome based on defects noticed in the gauge section of all the powder processed samples. The collective outcomes of this report provide further insight into the mechanisms by which carbides act as hydrogen traps; a means to process such carbides through powder metallurgy; and their associated mechanical performance in either a non-hydrogen or hydrogen-charged condition.

08 HYDROGEN↗