SEARCH · Engineering Papers
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.
An approach is disclosed that tracks memory transactions by a node. The approach establishes a transaction processing state corresponding to common virtual addresses accessed by a processing threads. Transactions are executed by the threads. A selected transaction is allowed to complete. In response to detecting a conflict in the transaction processing state, completion of a non-selected transaction is inhibited.
Since its launch in 2010, OpenACC has evolved into one of the most widely used portable programming models for accelerators on HPC systems today. Clacc is a project funded by the US Exascale Computing Project (ECP) to bring OpenACC support for C and C++ to the popular Clang and LLVM compiler infrastructure. In this paper, we describe Clacc's support for the OpenACC Profiling Interface, a critical component of the OpenACC specification that standardizes an interface that profiling tools and libraries can depend upon across OpenACC implementations. As part of Clacc's general strategy to build OpenACC support upon OpenMP, we describe how Clacc builds OpenACC Profiling Interface support upon an extended version of OMPT. We then describe how a major profiling and tracing toolkit within ECP, the TAU Performance System, takes advantage of this support. We also describe TAU's selective instrumentation support for OpenACC. Finally, using Clacc and TAU, we present example visualizations for several SPEC ACCEL OpenACC benchmarks running on an IBM AC922 node, and we show that the associated performance overhead is negligible.
This work presents a systematic methodology for the location selection of fast-charging stations for heavy-duty electric vehicles (EVs) based on both geospatial and electric grid analysis. The geospatial analysis is based on real-world geographic information system (GIS) data of road networks and existing supportive infrastructures. The grid analysis is implemented based on node-level analysis of potential impacts on voltages and power losses in the distribution system. A case study using a realistic, three-phase, unbalanced distribution feeder from California and extracted real-world GIS data is used to demonstrate the intuitiveness and effectiveness of the proposed methodology for the location selection of fast-charging stations for heavy-duty EVs considering both electric and existing transportation infrastructures.
We report herein the selective dimerization of propylene to branched hexenes using Ni-MFU-4$l$, a solid catalyst prepared by cation exchange. Analysis of the resulting product distribution demonstrates that the selectivity arises from 2,1-insertion and slow product reinsertion, mechanistic features reproduced by a molecular nickel tris-pyrazolylborate catalyst. Characterization of Ni-MFU-4$l$ by X-ray absorption spectroscopy provides evidence for discrete, tris-pyrazolylborate- like coordination of nickel, underscoring the small-molecule analogy that can be made at metal–organic framework nodes.
Today, multiGPU nodes are widely used in high-performance computing and data centers. However, current programming models do not provide simple, transparent, and portable support for automatically targeting multiple GPUs within a node on application areas of array programming. In this paper, we describe a new application programming interface based on the Kokkos programming model to enable array computation on multiple GPUs in a transparent and portable way across both NVIDIA and AMD GPUs. We implement different variations of this technique to accommodate the exchange of stencils (array boundaries) among different GPU memory spaces, and we provide autotuning to select the proper number of GPUs, depending on the computational cost of the operations to be computed on arrays, that is completely transparent to the programmer. We evaluate our multiGPU extension on Summit (#5 TOP500), with six NVIDIA V100 Volta GPUs per node, and Crusher that contains identical hardware/software as Frontier (#1 TOP500), with four AMD MI250X GPUs, each with 2 Graphics Compute Dies (GCDs)for a total of 8 GCDs per node. We also compare the performance of this solution against the use of MPI + Kokkos, which is the cur-rent de facto solution for multiple GPUs in Kokkos. Our evaluation shows that the new Kokkos solution provides good scalability for many GPUs and a faster and simpler solution (from a programming productivity perspective) than MPI + Kokkos.
We report that the accurate disease diagnosis and prognosis based on omics data rely on the effective identification of robust prognostic and diagnostic markers that reflect the states of the biological processes underlying the disease pathogenesis and progression. In this article, we present GCNCC, a Graph Convolutional Network-based approach for Clustering and Classification, that can identify highly effective and robust network-based disease markers. Based on a geometric deep learning framework, GCNCC learns deep network representations by integrating gene expression data with protein interaction data to identify highly reproducible markers with consistently accurate prediction performance across independent datasets possibly from different platforms. GCNCC identifies these markers by clustering the nodes in the protein interaction network based on latent similarity measures learned by the deep architecture of a graph convolutional network, followed by a supervised feature selection procedure that extracts clusters that are highly predictive of the disease state. By benchmarking GCNCC based on independent datasets from different diseases (psychiatric disorder and cancer) and different platforms (microarray and RNA-seq), we show that GCNCC outperforms other state-of-the-art methods in terms of accuracy and reproducibility.
The increase of computer processing speed is significantly outpacing improvements in network and storage bandwidth, leading to the big data challenge in modern science, where scientific applications can quickly generate much more data than that can be transferred and stored. As a result, big scientific data must be reduced by a few orders of magnitude while the accuracy of the reduced data needs to be guaranteed for further scientific explorations. Moreover, scientists are often interested in some specific spatial/temporal regions in their data, where higher accuracy is required. The locations of the regions requiring high accuracy can sometimes be prescribed based on application knowledge, while other times they must be estimated based on general spatial/temporal variation. In this paper, we develop a novel multilevel approach which allows users to impose region-wise compression error bounds. Our method utilizes the byproduct of a multilevel compressor to detect regions where details are rich and we provide the theoretical underpinning for region-wise error control. With spatially varying precision preservation, our approach can achieve significantly higher compression ratios than single-error bounded compression approaches and control errors in the regions of interest.We conduct the evaluations on two climate use cases – one targeting small-scale, node features and the other focusing on long, areal features. For both use cases, the locations of the features were unknown ahead of the compression. By selecting approximately 16% of the data based on multi-scale spatial variations and compressing those regions with smaller error tolerances than the rest, our approach improves the accuracy of post-analysis by approximately 2 × compared to single-error-bounded compression at the same compression ratio. Using the same error bound for the region of interest, our approach can achieve an increase of more than 50% in overall compression ratio.
This paper presents a decision-making framework based on an integrated artificial reasoning framework and Markov decision process (MDP). The integrated artificial reasoning framework provides a physics-based approach that converts system information into state transition models, and the analysis result will be represented by the transition probabilities that can be used with an MDP to find a traceable and explainable optimal pathway. A dynamic Bayesian network (DBN) is well suited for representing the structure of an MDP. The causality information among process variables (or among subsystems) is mathematically represented in a DBN by the conditional probabilities of the node’s states provided different probabilities of the parent node’s states. To define node states in a physically understandable manner, we used multilevel flow modeling (MFM). An MFM follows the fundamental energy and mass conservation laws and supports the selection of process variables that represent the system of interest so that causal relations among process variables are properly captured. An MFM-based DBN supports developing state transition models in an MDP to capture the effect of process variables of system having physical relations. The operators of the target system can capture stochastic system dynamics as multiple subsystem state transitions based on their physical relations and uncertainties coming from component degradation or random failures. We analyzed a simplified exemplary system to illustrate an optimal operational policy using the suggested approach.
The paper presents a system architecture for distributed sensing, networking and computing, its hardware implementation, and execution of initial flight experiments to validate theoretical findings. It induces development of distributed sensing requirements, framework, and architecture, development of distributed ground node hardware prototypes, integration of all nodes and testing of baseline functionalities, integration of in-house developed perception, migration and tracking software packages, establishing flight scenario and flyable path for a selected UAS, flying the air vehicle along the path, recording sensors measurements, pre-processing them and transferring the resulting data to an optimal computing center. It also addresses the challenges related to pre-flight hardware calibration, clock synchronization, sensor registration and establishing a communication network. Sensors data processing results demonstrate the functionality of the presented distributed architecture and satisfactory performance of the applied technologies.
The paper presents a system architecture for distributed sensing, networking and computing, its hardware implementation, and execution of initial flight experiments to validate theoretical findings. It induces development of distributed sensing requirements, framework, and architecture, development of distributed ground node hardware prototypes, integration of all nodes and testing of baseline functionalities, integration of in-house developed perception, migration and tracking software packages, establishing flight scenario and flyable path for a selected UAS, flying the air vehicle along the path, recording sensors measurements, pre-processing them and transferring the resulting data to an optimal computing center. It also addresses the challenges related to pre-flight hardware calibration, clock synchronization, sensor registration and establishing a communication network. Sensors data processing results demonstrate the functionality of the presented distributed architecture and satisfactory performance of the applied technologies.
Our nation's dependence on information networks makes it vital to anticipate disruptions or find weaknesses in these networks. But networks like the Internet are vast, distributed, and there is no mechanism to completely collect their structure. We are restricted to specific data collection schemes (like traceroute samples from router interfaces) that examine tiny portions of such a network. It has been empirically documented and theoretically proven that these measurements have significant biases, and direct inferences from them will be wrong. But these data collection mechanisms have limited flexibility and cannot be easily modified. Moreover, in many applications there are limits on how much data can be collected. How do we make accurate inferences of network properties with biased and limited measurements? The general problem this report deals with is how to work with incompletely observed networks. We will present several different approaches to this problem. First we will present an approach to estimate the degree distribution of a graph by sampling only a small portion of the vertices. This algorithm provides provably accurate results with sublinear samples. An alternative approach would be to try to enhance the information in the by selective collecting new information by probing for neighbors of a vertex or presence of individual edges. A different setting for working with incomplete arises when we have full access to local information, but do not have any global version of the graph. Can we still identify critical nodes in such a graph? We present an approach to identify such nodes efficiently. Finally, how can we put these ideas together to identify the structure of a network? We present an approach that can complement the existing approaches for network mapping. We start with an estimate of network structure based on existing network mapping methods. Then we find a critical router in the network, use the traffic through this network to selectively collect new data to enhance our prediction.
Cuprate superconductors host a multitude of low-energy optical phonons. Using time- and angle-resolved photoemission spectroscopy, we study coherent phonons in Bi 2 Sr 2 Ca 0.92 Y 0.08 Cu 2 O 8+δ . Sub-meV modulations of the electronic band structure are observed at frequencies of 3.94 ± 0.01 and 5.59 ± 0.06 THz. For the dominant mode at 3.94 THz, the amplitude of the band energy oscillation weakly increases as a function of momentum away from the node. Theoretical calculations allow identifying the observed modes as CuO 2 -derived A 1 g phonons. The Bi- and Sr-derived A 1 g modes which dominate Raman spectra in the relevant frequency range are absent in our measurements. Furthermore, this highlights the mode selectivity for phonons coupled to the near-Fermi-level electrons, which originate from CuO 2 planes and dictate thermodynamic properties.
We propose an efficient distributed out-of-memory implementation of the non-negative matrix factorization (NMF) algorithm for heterogeneous high-performance-computing systems. The proposed implementation is based on prior work on NMFk, which can perform automatic model selection and extract latent variables and patterns from data. In this work, we extend NMFk by adding support for dense and sparse matrix operation on multi-node, multi-GPU systems. The resulting algorithm is optimized for out-of-memory problems where the memory required to factorize a given matrix is greater than the available GPU memory. Memory complexity is reduced by batching/tiling strategies, and sparse and dense matrix operations are significantly accelerated with GPU cores (or tensor cores when available). Input/output latency associated with batch copies between host and device is hidden using CUDA streams to overlap data transfers and compute asynchronously, and latency associated with collective communications (both intra-node and inter-node) is reduced using optimized NVIDIA Collective Communication Library (NCCL) based communicators. Benchmark results show significant improvement, from 32X to 76x speedup, with the new implementation using GPUs over the CPU-based NMFk. Good weak scaling was demonstrated on up to 4096 multi-GPU cluster nodes with approximately 25,000 GPUs when decomposing a dense 340 Terabyte-size matrix and an 11 Exabyte-size sparse matrix of density 10 -6 .
High performance, parallel applications with irregular data accesses are becoming a critical workload class for modern systems. In particular, the execution of such workloads on emerging many-core systems is expected to be a significant component of applications in data mining, machine learning, scientific computing and graph analytics. However, power and energy constraints limit the capabilities of individual cores, memory hierarchy and on-chip interconnect of such systems, thus leading to architectural and software trade-os that must be understood in the context of the intended application’s behavior. Irregular applications are notoriously hard to optimize given their data-dependent access patterns, lack of structured locality and complex data structures and code patterns. We have ported two irregular applications, graph community detection using the Louvain method (Grappolo) and high-performance conjugate gradient (HPCCG), to the Tilera many-core system and have conducted a detailed study of platform-independent and platform-specific optimizations that improve their performance as well as reduce their overall energy consumption. To conduct this study, we employ an auto-tuning based approach that explores the optimization design space along three dimensions - memory layout schemes, GCC compiler flag choices and OpenMP loop scheduling options. We leverage MIT’s OpenTuner auto-tuning framework to explore and recommend energy optimal choices for different combinations of parameters. We then conduct an in-depth architectural characterization to understand the memory behavior of the selected workloads. Finally, we perform a correlation study to demonstrate the interplay between the hardware behavior and application characteristics. Using auto-tuning, we demonstrate whole-node energy savings and performance improvements of up to 49:6% and 60% relative to a baseline instantiation, and up to 31% and 45:4% relative to manually optimized variants.
Single non-heme Fe(II) ions present as structural moieties in several metal–organic frameworks (e.g., MIL-100, MIL-101, and MIL-808, where MIL indicates Materials of Institute Lavoisier) are identified by Kohn–Sham density functional calculations as promising catalysts for C–H bond activation, with energetic barriers as low as 40 kJ mol–1 for ethane and 60 kJ mol –1 for methane following the oxidative activation of iron. The rate-determining step is the activation of N 2 O and has a barrier of 140 kJ mol –1 . Through consideration of the full reaction profile leading to the corresponding alcohols, ethanol and methanol, we have identified key changes in the chemical composition of the node that would modulate catalytic activity. The thermal and chemical stabilities of these materials, together with the scalability of their syntheses, make them attractive catalysts for the selective low-temperature conversion of light alkanes to higher-value oxygenates.
Modeling of high data rate communication systems is different from the low data rate systems. Three simulations were built during the development phase of Carrier Sensed Multiple Access/Ring Network (CSMA/RN) modeling. The first was a model using SIMCRIPT based upon the determination and processing of each event at each node. The second simulation was developed in C based upon isolating the distinct object that can be identified as the ring, the message, the node, and the set of critical events. The third model further identified the basic network functionality by creating a single object, the node which includes the set of critical events which occur at the node. The ring structure is implicit in the node structure. This model was also built in C. Each model is discussed and their features compared. It should be stated that the language used was mainly selected by the model developer because of his past familiarity. Further the models were not built with the intent to compare either structure or language but because the complexity of the problem and initial results contained obvious errors, so alternative models were built to isolate, determine, and correct programming and modeling errors. The CSMA/RN protocol is discussed in sufficient detail to understand modeling complexities. Each model is described along with its features and problems. The models are compared and concluding observations and remarks are presented.
This dataset contains environmental monitoring data collected using solar-powered Multi-Function Research (MFR) Long Range Wide Area (LoRaWAN)-enabled nodes at 11 sites in Chicago, Illinois, as part of the DOE Urban Integrated Field Lab CROCUS project. The MFR node system consists of an Input/Output Digital Input Module (IB8) interface box (ICT International) providing wired connections for environmental sensors and an MFR-Node-L data logger that manages power, data processing, and LoRaWAN communication. The wireless data are ingested via Sage network (https://sagecontinuum.org/) nodes that contain LoRaWAN antennae. Measurements were collected from 11 MFR nodes deployed across Chicago State University (CSU), Northeastern Illinois University (NEIU), Northwestern University (NU), University of Illinois Chicago (UIC), West Woodlawn "Blacks in Green" (BIG), and Indian Boundary Prairies (IBP). Each MFR node supports a consistent suite of sensors measuring atmospheric, soil, and hydrological variables. Atmospheric measurements include 2m air temperature (°C), 2m vapor pressure deficit (kPa), and 2m shortwave/longwave radiation (incoming and outgoing, W/m²) measured using ATH-VPD and Apogee SN500 sensors. Soil measurements include volumetric water content (VWC, %) and temperature (°C) at four depths (15, 30, 45, and 60 cm below surface) using Meter Teros54 sensors, and heat flux (W/m²) at 10 cm depth using Huske HFP01-05 sensors. At selected locations, Meter Hydros21 sensors measure groundwater depth (mm), specific conductivity (dS/m), and temperature (°C). The dataset includes timestamps, site identifiers with location names, device IDs, Global Positioning System (GPS) coordinates, variable names with units, measurement depths, values, sensor names, and Sage node identifiers. All timestamps are in local Chicago time (CDT/CST). Quality control flags are provided using a 6-bit binary system indicating physical range violations, step spikes, 24-hour flat-line conditions, 6-hour jitter, 7-day ultra-low variance, and persistent high offset. Data is provided in CSV and CF-compliant NetCDF formats. This dataset is part of a larger collection of CROCUS environmental monitoring data, including linked datasets from Air Quality Transmitter (AQT) sensors, Weather Transmitter (WXT) sensors, and Sap Flow Meter (SFM1x) sensors.
A new metalloporphyrin framework of molybdenum Mo 2 O 4 (C 48 H 28 N 4 O 8 )·(CH 3 ) 2 NH·5H 2 O·2DMF (Mo 2 TCPP) was synthesized from tetrakis(4-carboxyphenyl)porphyrin (H 4 TCPP) and sodium molybdate dihydrate by a hydrothermal method. Mo 2 TCPP is a 3D network with two sub-units, in which both TCPP ligands and each Mo 2 dimer act as four connection nodes. The crystal structure was determined by single crystal analysis and further characterized by FTIR, SEM, EDX, PXRD, XPS and TGA. Here cumene hydrogen peroxide and hydrogen peroxide were used as oxidants to study the catalytic activity of new metalloporphyrins in the oxidation of cyclohexene at different temperatures. The conversion rate of cyclohexene and the selectivity of epoxycyclohexane were both higher than 99%, which was better than the previously published research results. The stability of the catalyst before and after the reaction was further tested for 10 runs without obvious degradation. The catalyst was stable in different solutions such as acidic, water and alkaline. These results shed light on the future development of new catalytic materials based on metalloporphyrin.