Search NASA⌕ Search

SEARCH · Search NASA

Results for “High Bandwidth Network”

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 433 records · Page 24

Non-resonant Bragg scattering four-wave mixing at near-visible wavelengths in low-confinement silicon nitride waveguides

Quantum state coherent frequency conversion processes—such as Bragg-scattering four-wave mixing (BSFWM)—hold promise as a flexible technique for networking heterogeneous and distant quantum systems. In this Letter, we demonstrate BSFWM within an extended (1.2-m) low-confinement silicon nitride waveguide and show that this system has the potential for near-unity frequency conversion in visible and near-visible wavelength ranges. Using sensitive classical heterodyne laser spectroscopy at low optical powers, we characterize the Kerr coefficient (∼1.55 W −1 m −1 ) and linear propagation loss (∼0.0175 dB/cm) of this non-resonant waveguide system, revealing a record-high nonlinear figure of merit (NFM = γ / α ≈ 3.85 W −1 ) for BSFWM of near-visible light in non-resonant silicon nitride waveguides. We predict how, at high yet achievable on-chip optical powers, this NFM would yield a comparatively large frequency conversion efficiency, opening the door to near-unity flexible frequency conversion without cavity enhancement and resulting bandwidth constraints.

Jaber, Nicholas (ORCID:000900070106809X)↗

Development, Manufacturing, and Preparation for Serial Production of Low Noise Seismometers (CRADA Final Report)

This was a collaborative effort between Lawrence Livermore National Security, LLC as manager and operator of Lawrence Livermore National Laboratory (LLNL) and Refraction Technology, Inc. (REF TEK), to collaborate on the development of a broadband, competitive low-noise seismometer, rugged and easy to use for field deployments. The work involved collaborative work between LLNL, REF TEK and a group led by Federal State Unitary Enterprise, Research Institute of Pulse Technique (RIPT), Moscow, Russia. The proposed work focused on bringing an improved version of the Russian SDSE seismometer from development phase to production in two versions. The first was a very low cost, rugged, broadband seismometer for field deployment that would achieve noise levels comparable to the standard earth low noise model (LNM) of the USGS. All three components were integrated into one case, and have sensitivity near 2000 v/m/s, and analog output with bandwidth of .01 to 40 Hz with high coherence. The second phase of development combined the results of previous instruments and these tests, and lowered the internal noise according to class A of standard Advanced National Seismic System (ANSS) within the band of 0.01 to 50 Hz for a dynamic range 143 - 128 dB. The second phase extended the development using improved electronics and modifications suggested by these tests of the earlier prototypes. These sensors would also be very competitive and aimed towards major improvements in their ruggedness compared to existing sensors from other manufacturers. Prototypes would undergo extensive testing at low-noise field sites for durability and functionality. The development would yield production prototypes capable of serial production at RPZ in Moscow. The collaboration between the Russian and US partners would improve the availability of rugged, broadband seismometers for commercial applications, research applications, seismic networks operated by our government and international organizations, and a spur collaboration on the development of additional types of seismometers.

Vergino, Eileen S.↗

Performance Prediction of Big Data Transfer Through Experimental Analysis and Machine Learning

Big data transfer in next-generation scientific ap- plications is now commonly carried out over connections with guaranteed bandwidth provisioned in High-performance Net- works (HPNs) through advance bandwidth reservation. To use HPN resources efficiently, provisioning agents need to carefully schedule data transfer requests and allocate appropriate band- widths. Such reserved bandwidths, if not fully utilized by the requesting user, could be simply wasted or cause extra overhead and complexity in management due to exclusive access. This calls for the capability of performance prediction to reserve bandwidth resources that match actual needs. Towards this goal, we employ machine learning algorithms to predict big data transfer perfor- mance based on extensive performance measurements, which are collected over a span of several years from a large number of data transfer tests using different protocols and toolkits between various end sites on several real-life physical or emulated HPN testbeds. We first identify a comprehensive list of attributes involved in a typical big data transfer process, including end host system configurations, network connection properties, and control parameters of data transfer methods. We then conduct an in-depth exploratory analysis of their impacts on application- level throughput, which provides insights into big data transfer performance and motivates the use of machine learning. We also investigate the applicability of machine learning algorithms and derive their general performance bounds for performance prediction of big data transfer in HPNs. Experimental results show that, with appropriate data preprocessing, the proposed machine learning-based approach achieves 95% or higher pre- diction accuracy in up to 90% of the cases with very noisy real-life performance measurements.

Yun, Daqing↗

GPU-Centric Communication on NVIDIA GPU Clusters with InfiniBand: A Case Study with OpenSHMEM

GPUs have become an essential component for building compute clusters with high compute density and high performance per watt. As such clusters scale to have 1000s of GPUs, efficiently moving data between the GPUs becomes imperative to get maximum performance. NVSHMEM is an implementation of the OpenSHMEM standard for NVIDIA GPU clusters which allows communication to be issued from inside GPU kernels. In earlier work, we have shown how NVSHMEM can be used to achieve better application performance on GPUs connected through PCIe or NVLink. As part of this effort, we implement IB verbs for Mellanox InfiniBand adapters in CUDA. We evaluate different design alternatives, taking into consideration the relaxed memory model, automatic memory access coalescing and thread hierarchy on the GPU. We also consider correctness issues that arise in these designs. We take advantage of these designs transparently or through API extensions in NVSHMEM. With micro-benchmarks, we show that a Nvidia Pascal P100 GPU is able saturate the network bandwidth using only one or two of its 56 available streaming multiprocessors (SM). On a single GPU using a single IB EDR adapter, we achieve a throughput of around 90 million messages per second. In addition, we implement a 2dstencil application kernel using NVSHMEM and compare its performance with a CUDA-aware MPI-based implementation that uses GPUDirect RDMA. Speedups in the range of 23% to 42% are seen for input sizes large enough to fill the occupancy of Nvidia Pascal P100 GPUs on 2 to 4 nodes indicating that there are gains to be had by eliminating the CPU from the communication path when all computation runs on the GPU.

Potluri, Sreeram↗

Bulk Data Mover

Bulk Data Mover (BDM) is a high-level data transfer management tool. BDM handles the issue of large variance in file sizes and a big portion of small files by managing the file transfers with optimized transfer queue and concurrency management algorithms. For example, climate simulation data sets are characterized by large volume of files with extreme variance in file sizes. The BDN achieves high performance using a variety of techniques, including multi-threaded concurrent transfer connections, data channel caching, load balancing over multiple transfer servers, and storage i/o pre-fetching. Logging information from the BDM is collected and analyzed to study the effectiveness of the transfer management algorithms. The BDM can accept a request composed of multiple files or an entire directory. The request also contains the target site and directory where the replicated files will reside. If a directory is provided at the source, then the BDM will replicate the structure of the source directory at the target site. The BDM is capable of transferring multiple files concurrently as well as using parallel TCP streams. The optimal level of concurrency or parallel streams depends on the bandwidth capacity of the storage systems at both ends of the transfer as well as achievable bandwidth of the wide-area network. Hardware req.-PC, MAC, Multi-platform & Workstation; Software req.: Compile/version-Java 1.50_x or above; Type of files: source code, executable modules, installation instructions other, user guide; URL: http://sdm.lbl.gov/bdm/

Natarajan, Vijaya↗

In-Containment Signal Conditioning and Transmission via Power Lines within High Dose Rate Areas of Nuclear Power Plants

Signal conditioning and transmission for sensor systems and networks within the containment of nuclear power plants (NPPs) still poses a challenge to engineers, particularly in the case of equipment upgrades for existing plants, temporary measurements, decommissioning of plants, but also for new builds. This paper presents an innovative method for efficient and cost-effective instrumentation within high dose rate areas inside the containment. A transmitter-receiver topology is proposed that allows simultaneous, unidirectional point-to-point transmission of multiple sensor signals by superimposing them on existing AC or DC power supply cables using power line communication (PLC) technology. Thereby the need for costly installation of additional cables and containment penetrations is eliminated. Based on commercial off-the-shelf (COTS) electronic parts, a radiation hard transmitter is designed to operate in harsh environment within the containment during full plant operation. Hardware modularity of the transmitter allows application specific tradeoffs between redundancy and channel bandwidth. At receiver side in non-radiated areas, signals are extracted from the power line, demodulated, and provided either in analog or digital output format. Laboratory qualification tests and field test results within a boiling water reactor (BWR) are validating the proof of concept of the proposed system. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

A High-Current Pulsed Prototype Power Supply

The Accelerator Controls Operations Research Network (ACORN) project aims to modernize the accelerator control system and replace aging power supplies at Fermilab. As part of this effort, outdated RF ferrite bias power supplies will be redesigned. These power supplies are essential for tuning the resonant frequency of RF cavities by delivering programmable current outputs of up to 2500 A and voltages ranging from −10 V to +35 V. They operate at a repetition rate of 15 Hz in the Booster ring, and 1 Hz at the Main Injector ring. The power supplies utilize a bank of transistors in the linear region, connected in parallel with the load, to actively regulate the output current from a 12-pulse SCR bridge. To support this upgrade, a new bias power supply topology was developed as proof of concept. The design utilizes an IGBT Hbridge operating in Pulse Width Modulation (PWM) mode, controlled by a microcontroller. A prototype, constructed using spare components, successfully delivered an output current of 500 A at a repetition rate of 15 Hz during initial testing. The circuit's bandwidth was measured at 480 Hz, highlighting opportunities for further optimization in the controller design to achieve the target bandwidth of 2 kHz.

Bullman, Austin [ORNL]↗

Climate Science Performance, Data and Productivity on Titan

Climate Science models are flagship codes for the largest of high performance computing (HPC) resources, both in visibility, with the newly launched Department of Energy (DOE) Accelerated Climate Model for Energy (ACME) effort, and in terms of significant fractions of system usage. The performance of the DOE ACME model is captured with application level timers and examined through a sizeable run archive. Performance and variability of compute, queue time and ancillary services are examined. As Climate Science advances in the use of HPC resources there has been an increase in the required human and data systems to achieve programs goals. A description of current workflow processes (hardware, software, human) and planned automation of the workflow, along with historical and projected data in motion and at rest data usage, are detailed. The combination of these two topics motivates a description of future systems requirements for DOE Climate Modeling efforts, focusing on the growth of data storage and network and disk bandwidth required to handle data at an acceptable rate.

Mayer, Benjamin W↗

Evaluating the Potential of Cray Gemini Interconnect for PGAS Communication Runtime Systems

The Cray Gemini Interconnect has been recently introduced as the next generation network for building scalable multi-petascale supercomputers. The Cray XE6 systems, which use the Gemini Interconnect are becoming available with Message Passing Interface (MPI) and Partitioned Global Address Space (PGAS) Models such as as Global Arrays, Unified Parallel C, Co-Array Fortran and Cascade High Performance Language. These PGAS models use one-sided communication runtime systems such as MPI-Remote Memory Access, Aggregate Remote Memory Copy Interface and proprietary communication runtime systems. The primary objective of our work is to study the potential of Cray Gemini Interconnect by designing application specific micro-benchmarks using the DMAPP userspace library provided by the Gemini Interconnect. To meet this objective, we design micro-benchmarks to study the performance of simple communication primitives and application specific micro-benchmarks to understand the behavior of Gemini Interconnect at scale. In our experiments, the Gemini Interconnect can achieve a peak bandwidth of 6911 MB/s and a latency of 1us for get communication primitive. Scalability tests for atomic memory operations and shift communication operation up to 65536 processes shows the efficacy of the Gemini Interconnect.

Vishnu, Abhinav↗

Scalable and Fault Tolerant Failure Detection and Consensus

Future extreme-scale high-performance computing systems will be required to work under frequent component failures. The MPI Forum's User Level Failure Mitigation proposal has introduced an operation, MPI_Comm_shrink, to synchronize the alive processes on the list of failed processes, so that applications can continue to execute even in the presence of failures by adopting algorithm-based fault tolerance techniques. This MPI_Comm_shrink operation requires a fault tolerant failure detection and consensus algorithm. This paper presents and compares two novel failure detection and consensus algorithms. The proposed algorithms are based on Gossip protocols and are inherently fault-tolerant and scalable. The proposed algorithms were implemented and tested using the Extreme-scale Simulator. The results show that in both algorithms the number of Gossip cycles to achieve global consensus scales logarithmically with system size. The second algorithm also shows better scalability in terms of memory and network bandwidth usage and a perfect synchronization in achieving global consensus.

Katti, Amogh↗

On-chip probabilistic inference for charged-particle tracking at the sensor edge

Modern scientific instruments operate under increasingly extreme constraints on bandwidth, latency, and power. Inference at the sensor edge determines experimental data collection efficiency by deciding which information to save for further analysis. Particle tracking detectors at the Large Hadron Collider exemplify this challenge: pixelated silicon sensors generate rich spatiotemporal ionization patterns, yet most of this information is discarded due to data-rate limitations. Concurrently, advancements in co-design tools provide rapid turn-around for incorporating machine learning into application-specific integrated circuits, motivating designs for particle detectors with new integrated technologies. We demonstrate that neural networks embedded in the front-end electronics can infer charged-particle kinematic parameters from a single silicon layer. We regress hit positions and incident angles with calibrated uncertainties, while satisfying stringent constraints on numerical precision, latency, and silicon area. Our results establish a path toward probabilistic inference directly at the edge, opening new opportunities for intelligent sensing in high-rate scientific instruments.

Das, Arghya Ranjan [Purdue U.] (ORCID:000000018451↗

The future of time and frequency dissemination

I will try to extrapolate the changes in the dissemination of time and frequency information that have taken place during the last 25 years to predict the future developments both in the methods of disseminating time and frequency and in the kinds of customers we will be asked to serve. Two important developments are likely to play pivotal roles in driving the evolution of dissemination. The first is the commercial availability of very high quality clocks -- devices whose performance may eventually rival that of the current generation of primary frequency standards. The widespread use of these devices may blur the traditional distinction between client and server, and may replace it with a more symmetrical interchange of data among peers. The second is the increasing demand for digital time and frequency information driven by the increasing sophistication of everything from traffic lights to electric power meters. The needs of these individual users may not tax the state of the art of primary frequency standards in principle, but their large numbers and wide geographical distribution present a technological challenge that is difficult to meet at a reasonable price using existing methods. Some of these problems may be solved (or at least addressed) using developments in communications and consumer electronics such as the increasing use of fiber-optic telephone circuits and the increasing bandwidth and sophistication of the cable network used to transmit television pictures. To be useful, these advances in hardware must stimulate parallel advances in software algorithms and methods. These advances are more difficult to predict with great confidence, but the developments of the last few years will be examined to provide some indications of the future.

Levine, Judah↗

Dynamic Shared Computing Resources for Multi-Robot Mars Exploration

The NASA roadmap for 2020 and beyond includes several key technologies which will have a game-changing impact on planetary exploration. The first of these is High Performance Spaceflight Computing (HPSC), which will provide orders of magnitude increases in processing power for next-generation rovers and orbiters (Doyle et al. 2013). The second is Delay Tolerant Networking, which overlays the Deep Space Network, providing internet-like abstractions and store-forward to route data through intermittent delays in connectivity. The third is a trend toward small, co-dependent robots included in flagship missions (MarCO, PUFFER, and Mars Heli). Taken together, these imply an increasing amount of communication and computing heterogeneity on Mars in coming decades. Motivated by these technological trends, we study the concept of Mars on-site shared analysis, information, and communication (MOSAIC) for Mars exploration. The key algorithmic problem associated with MOSAIC networks is simultaneous scheduling of computation, communication, and caching of data, which we illustrate using the three scenarios. We present models, preliminary solutions, and simulation results for two scenarios, showing how mission efficiency relates to communication bandwidth, processing power, geography of the environment, and optimal scheduling of computation, communication, and data caching. The third scenario illustrates future directions of this work.

Chien, Steve↗

Graph Analytics on Jellyfish topology

Because large unstructured datasets is important for many science domains, distributed graph analytics is critical to many scientists. Unfortunately, obtaining scaling and performance for irregular communication is challenging because contemporary network interconnects are primarily designed to maximize bandwidths of fixed-neighborhoods large-message exchanges (e.g., stencils). Although there is no consensus on the “best” network topologies for irregular communication, unstructured graph-based interconnects can be more suitable. We analyze three popular graph workloads – clustering, pattern enumeration, and traversal — on comparable networks (in terms of resources and costs) constructed from Jellyfish Random Regular, Dragonfly and Fat tree topologies, varying the routing algorithms. Using packet-level simulations, we demonstrate up to 60% improvement in communication time with Jellyfish due to diversity of the short paths between arbitrary endpoints, which can reduce overall network stalls and congestion.

Graph Analytics, network topology, interconnect, H↗

Benefits of Ka-band GaN MMIC High Power Amplifiers With Wide Bandwidth and High Spectral/Power Added Efficiencies for Cognitive Radio Platforms

A cognitive radio on a future NASA near-Earth spacecraft will be capable of sensing its environment and dynamically adapting its operating parameters to provide the desired SATCOM service to the mission. A key component that can enable this type of operation is a high-power amplifier (HPA) that resides on the radio platform. In this paper, we present the RF performance characteristics of a Ka-band gallium nitride (GaN) monolithic microwave integrated circuit (MMIC) based HPA for cognitive radio platforms. These characteristics include the output power, gain, power added efficiency (PAE), RMS error vector magnitude (EVM), spectral efficiency, 3rd-order intermodulation distortion (IMD) products, spectrum, spectral regrowth, noise figure (NF), and phase noise. The data presented indicates that the HPA meets NTIA, military, and commercial spectral mask requirements. In addition, we discuss the benefits offered by the above performance characteristics toward the design and implementation of a cognitive radio platform. Furthermore, as examples, we discuss three potential use cases that apply artificial intelligence (AI) and machine learning (ML) techniques and exploit the performance characteristics discussed above to provide a knowledge-based cognitive radio platform design for SATCOM. Thus, cognitive radios with performance flexibility can enable roaming and provide seamless interoperability autonomously in the future between NASA, commercial, and other space networks owned by U.S. government agencies.

Gallium nitride↗

Benefits of Ka-band GaN MMIC High Power Amplifiers With Wide Bandwidth and High Spectral/Power Added Efficiencies for Cognitive Radio Platforms

A cognitive radio on a future NASA near-Earth spacecraft will be capable of sensing its environment and dynamically adapting its operating parameters to provide the desired SATCOM service to the mission. A key component that can enable this type of operation is a high-power amplifier (HPA) that resides on the radio platform. In this report, we present the RF performance characteristics of a Ka-band gallium nitride (GaN) monolithic microwave integrated circuit (MMIC) based HPA for cognitive radio platforms. These characteristics include the output power, gain, power added efficiency (PAE), RMS error vector magnitude (EVM), spectral efficiency, 3rdorder intermodulation distortion (IMD) products, spectrum, spectral regrowth, noise figure (NF), phase noise, and group delay. The data presented indicates that the HPA meets NTIA, military, and commercial spectral mask requirements. In addition, we discuss the benefits offered by the above performance characteristics toward the design and implementation of a cognitive radio platform. Furthermore, as examples, we discuss three potential use cases that apply artificial intelligence (AI) and machine learning (ML) techniques and exploit the performance characteristics discussed above to provide a knowledge-based cognitive radio platform design for SATCOM. Thus, cognitive radios with performance flexibility can enable roaming and provide seamless interoperability autonomously in the future between NASA, commercial, and other space networks owned by U.S. government agencies.

Gallium nitride↗

Distributed Power Flow Control: Distributed Power Flow Control using Smart Wires for Energy Routing

GENI Project: Smart Wire Grid is developing a solution for controlling power flow within the electric grid to better manage unused and overall transmission capacity. The 300,000 miles of high-voltage transmission line in the U.S. today are congested and inefficient, with only around 50% of all transmission capacity utilized at any given time. Increased consumer demand should be met in part with more efficient and an economical power flow. Smart Wire Grid’s devices clamp onto existing transmission lines and control the flow of power within—much like how internet routers help allocate bandwidth throughout the web. Smart wires could support greater use of renewable energy by providing more consistent control over how that energy is routed within the grid on a real-time basis. This would lessen the concerns surrounding the grid’s inability to effectively store intermittent energy from renewables for later use.

ARPA-E↗