Search NASA⌕ Search

SEARCH · Search NASA

Results for “data access”

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 1,747 records · Page 97

Method for Accessing Distributed Heterogeneous Databases

A scenario of relational, hierarchial, and network data bases is presented and a distributed access view integrated data base system (DAVID) is described for uniformly accessing data bases which are heterogeneous and physically distributed. The DAVID system is based on data base logic so that the relational approach is generalized to the heterogeneous approach. The global data manager is explained as are global data manipulation languages which can operate on all the data bases and can query the data dictionary and the data directory.

Jacobs, B. E.↗

Status Report on Transfer of Physical and Hydraulic Properties Databases to the Hanford Environmental Information System - PNNL Remediation Decision Support Project, Task 1, Activity 6

This document provides a status report on efforts to transfer physical and hydraulic property data from PNNL to CHPRC for incorporation into HEIS. The Remediation Decision Support (RDS) Project is managed by Pacific Northwest National Laboratory (PNNL) to support Hanford Site waste management and remedial action decisions by the U.S. Department of Energy and their contractors. The objective of Task 1, Activity 6 of the RDS project is to compile all available physical and hydraulic property data for sediments from the Hanford Site, to port these data into the Hanford Environmental Information System (HEIS), and to make the data web-accessible to anyone on the Hanford Local Area Network via the so-called Virtual Library. These physical and hydraulic property data are used to estimate parameters for analytical and numerical flow and transport models that are used for site risk assessments and evaluation of remedial action alternatives. In past years efforts were made by RDS project staff to compile all available physical and hydraulic property data for Hanford sediments and to transfer these data into SoilVision®, a commercial geotechnical software package designed for storing, analyzing, and manipulating soils data. Although SoilVision® has proven to be useful, its access and use restrictions have been recognized as a limitation to the effective use of the physical and hydraulic property databases by the broader group of potential users involved in Hanford waste site issues. In order to make these data more widely available and useable, a decision was made to port them to HEIS and to make them web-accessible via a Virtual Library module. In FY08 the original objectives of this activity on the RDS project were to: (1) ensure traceability and defensibility of all physical and hydraulic property data currently residing in the SoilVision® database maintained by PNNL, (2) transfer the physical and hydraulic property data from the Microsoft Access database files used by SoilVision® into HEIS, which is currently being maintained by CH2M-Hill Plateau Remediation Company (CHRPC), (3) develop a Virtual Library module for accessing these data from HEIS, and (4) write a User's Manual for the Virtual Library module. The intent of these activities is to make the available physical and hydraulic property data more readily accessible and useable by technical staff and operable unit managers involved in waste site assessments and remedial action decisions for Hanford. In FY08 communications were established between PNNL and staff from Fluor-Hanford Co. (who formerly managed HEIS) to outline the design of a Virtual Library module that could be used to access the physical and hydraulic property data that are to be transferred into HEIS. Data dictionaries used by SoilVision® were also provided to Fluor-Hanford personnel who are now with CHPRC. During ongoing work to ensure traceability and defensibility of all physical and hydraulic property data that currently reside in the SoilVision® database, it was recognized that further work would be required in this effort before the data were actually ported into HEIS. Therefore work on the Virtual Library module development and an accompanying User's Guide was deferred until an unspecified later date. In FY09 efforts have continued to verify the traceability and defensibility of the physical and hydraulic property datasets that are currently being maintained by PNNL. Although this is a work in progress, several of these datasets are now ready for transfer to CHRPC for inclusion in HEIS. The actual loading of data into HEIS is performed by CHPRC staff, so after the data are transferred from PNNL to CHPRC, it will be the responsibility of CHPRC to ensure that these data are loaded and made accessible. This document provides a status report on efforts to transfer physical and hydraulic property data from PNNL to CHPRC for incorporation into HEIS.

54 ENVIRONMENTAL SCIENCES↗

Guided Tour of Pythonian Museum

At http:hdfeos.orgzoo, we have a large collection of Python examples of dealing with NASA HDF (Hierarchical Data Format) products. During this hands-on Python tutorial session, we'll present a few common hacks to access and visualize local NASA HDF data. We'll also cover how to access remote data served by OPeNDAP (Open-source Project for a Network Data Access Protocol). As a glue language, we will demonstrate how you can use Python for your data workflow - from searching data to analyzing data with machine learning.

hdf↗

Maintenance of Maywood Laboratory Operations Support During Contractor Transition and Replacement of the Project Databases - 16329

In December 2013, Cabrera Services (Cabrera) executed transition of the FUSRAP Maywood Superfund Site remediation project operations, including an on-site accredited environmental lab, from a large business that had established technology infrastructure in place for over a decade which required updating to support the expected remediation scope and client needs. As of November 2015, Cabrera has shipped over 116,000 yd{sup 3} of contaminated soil to disposal facilities. During the transition, the USACE FUSRAP Maywood Laboratory (UFML) continued to support project operations, providing the expected timely and reliable results and maintaining their New Jersey Department of Environmental Protection certification for all parameters of concern, while having to cope with new challenges and additional tasks having the potential to disrupt the laboratory's ability to support project operations. This case study examines the challenges with maintaining a certified radiochemistry laboratory and an aggressive remediation schedule during a period of prime contractor transition which included a complete IT infrastructure overhaul, and conversion of the existing Microsoft Access laboratory database and project Electronic Data Management System (EDMS) to a customized commercial solution. Challenges: I. Procurement of a new database subcontractor and working closely with them to achieve the following: a. Migration of all environmental (including radiological, chemical, and quality control), health and safety, and spatial data as well as reports, queries, and analytical instrument software processes from a MS Access data management system to a SQL server system; and b. Verification that the functionality and data integrity of the databases remained intact. II. The onsite laboratory was changed over to a new onsite network and the following tasks were completed with little or no interruption to site operations by January 10, 2014: i. Installation of new instrument personal computers (PCs) and making them fully operational; ii. Configuration of the lab domain and server to host the LIMS (replaced all servers/domain/workstations with Windows 7 and Server 2012 machines). iii. Networking of the laboratory building into the rest of the project site via fiber optic cabling across the site. iv. Upgrade of the wide area network (WAN) data connection to high-speed Fiber Optic service; v. Upgrade of the telephones into voice over internet protocol (VOIP) phones. vi. Splitting of the database into front ends and a centralized back-end hosted on the new server to allow simultaneous access for fixes and improvements. III. Phone service interruption; phone systems were operational by January 10, 2014. Additional Tasks: I. Revision of the USACE FUSRAP Maywood Laboratory (UFML) Quality Systems Manual Rev. 1 Submitted 01/16/14; II. Writing of the Uniform Federal Policy Quality Assurance Project Plan Rev. 1 (UFP-QAPP) submitted 1/21/14. This document replaced the prior project data quality plan, the Chemical Data Quality Management Plan. III. Support of USACE communications to the New Jersey Department of Environmental Protection (NJDEP) on lab management and personnel transition to Cabrera. This work was important since it ensured a smooth transition and shows how critical project records infrastructure such as databases, can be upgraded with minimal interruption of project support. Such work may provide benefit to other similar projects, especially long term projects where database infrastructure upgrades may be beneficial in terms of improved efficiencies in data recording and transfer, and improved security in data storage. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Electric Vehicle Charging Analytics and Reporting Tool (EV-ChART): Data Format and Preparation Guidance (V.5.0)

The Joint Office of Energy and Transportation maintains the Electric Vehicle Charging Analytics and Reporting Tool (EV-ChART), which provides a centralized hub for submitting electric vehicle (EV) charging infrastructure data directed by the Federal Highway Administration (23 CFR 680.112(a)-(c)). EV-ChART provides a streamlined data submission process and an integrated set of analytic tools, connects to other data sources, and empowers data sharing and access across stakeholders, including the public. Any data shared publicly will be aggregated and anonymized to stay in accordance with 23 CFR 680. This EV-ChART Data Format and Preparation Guidance provides a comprehensive overview of the data reporting requirements as authorized under 23 CFR 680.112(a)-(c)). The guidance is intended to be used alongside the EV-ChART Data Input Template, which defines the tabular data structure that these data submissions must follow. Per 23 CFR 680.112(a)-(c), the annual and quarterly data submissions are required of all National Electric Vehicle Infrastructure (NEVI) Formula Program projects, as well as projects for the construction of publicly accessible EV chargers that are funded with funds made available under Title 23, United States Code, including any EV charging infrastructure project funded with federal funds that is treated as a project on a federal-aid highway. One-time data submissions are required of both the NEVI Formula Program projects and grants awarded under 23 U.S.C. 151(f) for projects that are for EV charging stations located along and designed to serve the users of designated Alternative Fuel Corridors (AFCs). Other information and data required in 23 CFR 680, such as 23 CFR 680.112(d), 23 CFR 680.116(c), and 23 CFR 680.106(a), are not discussed in this guidance.

33 ADVANCED PROPULSION SYSTEMS↗

The RCSB Protein Data Bank: new resources for research and education

The Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB) develops tools and resources that provide a structural view of biology for research and education. The RCSB PDB web site (http://www.rcsb.org) uses the curated 3D macromolecular data contained in the PDB archive to offer unique methods to access, report and visualize data. Recent activities have focused on improving methods for simple and complex searches of PDB data, creating specialized access to chemical component data and providing domain-based structural alignments. New educational resources are offered at the PDB-101 educational view of the main web site such as Author Profiles that display a researcher’s PDB entries in a timeline. To promote different kinds of access to the RCSB PDB, Web Services have been expanded, and an RCSB PDB Mobile application for the iPhone/iPad has been released. These improvements enable new opportunities for analyzing and understanding structure data.

59 BASIC BIOLOGICAL SCIENCES↗

Open Science for Plants in Space: Data Sharing, Standards, and Informatics for Reuse and Knowledge Discovery

Upcoming deep space missions rely on plants and crops for crew and ecosystem health. Access to space plant data enables scientists to gain a deeper understanding of biological responses to ionizing radiation, altered gravity, low atmospheric pressure, elevated CO2, and altered photoperiods. Open Science is the practice of making research available to all, while respecting diverse cultures, fostering collaborations with equity. 2023 is the ‘Year of Open Science’, and NASA has a 5-year Transform to Open Science (TOPS) mission designed to rapidly transform the agency toward an inclusive culture of open science. NASA’s Open Science Data Repository (OSDR) developed by NASA’s Biological and Physical Sciences Division provides access to data from space-relevant biological experiments. OSDR combines two databases, GeneLab and Ames Life Sciences Data Archive (ALSDA) to maximize access to standardized ‘omics (e.g., transcriptomics, proteomics) and phenotypic data (e.g., microscopy, biomass), respectively. OSDR started in 2014 with the creation of the first space-relevant FAIR (Findable, Accessible, Interoperable, Reusable) biological ‘omics repository (GeneLab), providing detailed metadata on investigation, sample, and assay levels. Today, GeneLab hosts 62 plant datasets which have led to 5 published peer-reviewed meta-analysis publications. Most of these publications were collaboration efforts under the OSDR Analysis Working Groups (AWGs). AWGs provide great opportunities for investigators to collaborate and set new standards for space-relevant data and metadata. The AWGs are welcoming any ASPB members interested in providing plant expertise for space biology. The addition of ALSDA to OSDR is also expanding analysis capability beyond ‘omics. Now is the time to get involved as a Subject Matter Expert as we establish the framework for modern plant data archiving through the AWGs. Investigators are invited to submit their space-relevant plant datasets to OSDR and visit the site to learn about the tools OSDR has to offer (osdr.nasa.gov/bio).

FAIR↗

RDMA vs. RPC for Implementing Distributed Data Structures

Distributed data structures are key to implementing scalable applications for scientific simulations and data analysis. In this paper we look at two implementation styles for distributed data structures: remote direct memory access (RDMA) and remote procedure call (RPC). We focus on operations that require individual accesses to remote portions of a distributed data structure, e.g., accessing a hash table bucket or distributed queue, rather than global operations in which all processors collectively exchange information. We look at the trade-offs between the two styles through microbenchmarks and a performance model that approximates the cost of each. The RDMA operations have direct hardware support in the network and therefore lower latency and overhead, while the RPC operations are more expressive but higher cost and can suffer from lack of attentiveness from the remote side. We also run experiments to compare the real-world performance of RDMA- and RPC-based data structure operations with the predicted performance to evaluate the accuracy of our model, and show that while the model does not always precisely predict running time, it allows us to choose the best implementation in the examples shown. We believe this analysis will assist developers in designing data structures that will perform well on current network architectures, as well as network architects in providing better support for this class of distributed data structures.

distributed data structures↗

Dynamic object management for distributed data structures

In distributed-memory multiprocessors, remote memory accesses incur larger delays than local accesses. Hence, insightful allocation and access of distributed data can yield substantial performance gains. The authors argue for the use of dynamic data management policies encapsulated within individual distributed data structures. Distributed data structures offer performance, flexibility, abstraction, and system independence. This approach is supported by data from a trace-driven simulation study of parallel scientific benchmarks. Experimental data on memory locality, message count, message volume, and communication delay suggest that data-structure-specific data management is superior to a single, system-imposed policy.

Totty, Brian K.↗

RadLab and the Environmental Data Application Dashboard: Graphical and Programming Interfaces for Interrogation of Space Telemetry Data

Sensors on the International Space Station (ISS) and multiple spacecraft elsewhere in Earth orbit and in deep space continuously monitor and collect environmental data, transmitting this information back to Earth. These data include ionizing radiation and, on the ISS, CO2, relative humidity levels, and temperature, and are of great importance to space biology research. Ionizing radiation in particular has been established in ground-based experiments as being correlated with increased risk of carcinogenesis and cardiovascular and neurological effects. Looking ahead to future long duration crewed missions beyond low Earth orbit, the ability to study how factors including CO2 levels, light cycle, temperature modulate the response to ionizing radiation and microgravity is essential. To date, access to these data has been fragmented across space agencies, spacecraft, and databases. To address this issue, NASA’s Open Science Data Repository (osdr.nasa.gov) has developed two Web applications: the Environmental Data Application (EDA) and a radiation-specific RadLab. Each consists of an API (application programming interface) and an associated GUI (graphical user interface) that provide single points of access to the data. To date, OSDR has focused on the sensors from payloads and radiation detectors located on the ISS. The Web applications process telemetry information and associated data, such as spacecraft location and orientation, from multiple international databases. The applications’ request syntax enables users to interrogate these data by craft, sensor type, time range, radiation type (galactic cosmic rays, solar particle events, the contribution of the South Atlantic Anomaly), facilitating arbitrary comparisons of original source data at varying time resolutions. The applications provide programmatic access for use in computational pipelines and GUIs for data visualization and exploration, making these data FAIR (Findable, Accessible, Interoperable, and Reusable), complementing the biological data contained in OSDR, and providing the space science community with a valuable resource for scientific analyses.

radiation↗

OpenSolar: Promoting the Openness and Accessibility of Diverse Public Solar Datasets

Observational solar data is the foundation of data-driven research in solar power grid integration and power system operations. Compared to other fields in data science, the openness and accessibility of solar data is lacking, which prevents solar data science from catching up with the emerging trends of data science (e.g., deep learning). In this paper, OpenSolar, a package with both R and Python versions, is developed to enhance the openness and accessibility of publicly available solar datasets. The OpenSolar package provides access to multiple types of solar data, primarily from four datasets: (1) the National Renewable Energy Laboratory (NREL) Solar Power Data for Integration Studies dataset, (2) the NREL Solar Radiation Research Laboratory dataset, (3) the Sheffield Solar-Microgen database, and (4) the Dataport database. Unlike other open solar datasets that only contain meteorological data, the four datasets in the OpenSolar package also contain behind-the-meter power data, sky images, and solar power data with satisfactory temporal and spatial resolution and coverage. The overview, quality-control methods, and potential usage of the datasets, in conjunction with sample code implementing the OpenSolar functions, are described in this paper. The package is expected to assist in bridging the gaps between the research fields of solar energy, power systems, and data science.

14 SOLAR ENERGY↗

TRUST, Trustworthiness and EOSDIS

In recent years there has been considerable attention by the international scientific research and applications community to ensure high quality of data and information management. The terms FAIR (Findable, Accessible, Interoperable, Reusable) data, TRUST (Transparency, Responsibility, User Community, Sustainability, and Technology) principles, and CARE (Collective Benefit, Authority to Control, Responsibility, and Ethics) principles have come into vogue during the last decade. NASA has been managing data and information for over 60 years. NASA’s Earth Observing System Data and Information System (EOSDIS) has been in operation for over 25 years, managing most of NASA’s Earth science data. Trustworthiness is a goal that NASA has always strived to achieve or exceed, because it: enables the success of any NASA science mission; inspires general science research and applications; justifies the cost of operations; contributes to the value of NASA’s Open Data Policy; and influences the long term, historical view for the data collection. Given the recent growth of interest in TRUST principles, it is useful to assess and show how NASA’s attention to trustworthiness maps into those principles. This presentation addresses shows how the various steps that have been taken by the Earth Science Data and Information System (ESDIS) Project in the implementation and evolution of EOSDIS map into the TRUST principles.

Remote Sensing↗

UMap: An application-oriented user level memory mapping library

Exploiting the prominent role of complex memories in exascale node architecture, the UMap page fault handler offers new capabilities to access large memory-mapped data sets directly. UMap provides flexible configuration options to customize page handling to each application, including analysis of massive observational and simulation data sets. The high-performance design features I/O decoupling, dynamic load balancing, and application-level controls. Page faults triggered by application threads and processes accessing data mapped to a UMapp’ed region are handled via the Linux userfaultfd protocol, an asynchronous message-oriented kernel-user communication mechanism that avoids the context switch penalty of traditional signal fault handlers. UMap is fully open source. In this paper, we give an overview of the UMap library architecture, its extensible plugin architecture, and the use/performance of UMap in emerging heterogeneous memory hierarchies such as near-node Non-volatile Memory (NVM) and network attached memories. We highlight new capabilities in two pagefault management plugins, the NetworkStore and SparseStore. We demonstrate the integration between UMap and multiple ECP products including Caliper, Metall, ZFP, Mochi, and Ripples.

97 MATHEMATICS AND COMPUTING↗

Development of a phonon-based sampling method for thermal neutron scattering data

Simulations of reactor systems require access to accurate nuclear data. For many systems, thermal neutron scattering data can have large effects on the eigenvalue and neutron flux distributions. Inelastic thermal neutron scattering can excite or de-excite vibrational, rotational, and translational modes in a material, so thermal scattering evaluations are often obtained by summing over the number of phonons created/destroyed by a scattering event. In recent years, the thermal scattering cross sections and angular distributions have greatly improved in accuracy, but the format in which this data is delivered to simulation codes has remained virtually unchanged. Thermal scattering data is typically either compiled into large tables and sorted by incoming neutron energy, outgoing neutron energy, scattering angle, and material temperature, or represented as cumulative distribution functions of momentum exchange or energy exchange. Either method can be quite memory intensive when fine bins are used. In an effort to decrease the amount of space that processed thermal scattering data requires, an alternate format is proposed. The phonon-based sampling method introduced here can sample the number of phonons excited for each collision, the change in neutron energy, and the scattering angle while avoiding pre-computed angular bins and limiting the amount of data that is dependent on incoming energy. Through this method, the generation and storage of large interpolation tables is avoided, which could have benefits in both memory storage and accuracy. While the initial implementation of this method is slower than current alternatives, it is significantly more resistant to grid coarseness errors and has good potential for improvement. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Distributed caching for processing raw arrays

As applications continue to generate multi-dimensional data at exponentially increasing rates, fast analytics to extract meaningful results is becoming extremely important. The database community has developed array databases that alleviate this problem through a series of techniques. In-situ mechanisms provide direct access to raw data in the original format---without loading and partitioning. Parallel processing scales to the largest datasets. In-memory caching reduces latency when the same data are accessed across a workload of queries. However, we are not aware of any work on distributed caching of multi-dimensional raw arrays. In this paper, we introduce a distributed framework for cost-based caching of multi-dimensional arrays in native format. Given a set of files that contain portions of an array and an online query workload, the framework computes an effective caching plan in two stages. First, the plan identifies the cells to be cached locally from each of the input files by continuously refining an evolving R-tree index. In the second stage, an optimal assignment of cells to nodes that collocates dependent cells in order to minimize the overall data transfer is determined. We design cache eviction and placement heuristic algorithms that consider the historical query workload. A thorough experimental evaluation over two real datasets in three file formats confirms the superiority -- by as much as two orders of magnitude -- of the proposed framework over existing techniques in terms of cache overhead and workload execution time.

Zhao, Weijie↗

An infrared, Raman, and X-ray database of battery interphase components

Further improvements to lithium-ion and emerging battery technologies can be enabled by an improved understanding of the chemistry and working mechanisms of interphases that form at electrochemically active battery interfaces. However, it is difficult to collect and interpret spectra of interphases for several reasons, including the presence of a variety of compounds. To address this challenge, we herein present a vibrational spectroscopy and X-ray diffraction data library of ten compounds that have been identified as interphase constituents in lithium-ion or emerging battery chemistries. The data library includes attenuated total reflectance Fourier transform infrared spectroscopy, Raman spectroscopy, and X-ray diffraction data, collected in inert atmospheres provided by custom sample chambers. The data library presented in this work (and online repository) simplifies access to reference data that is otherwise either diffusely spread throughout the literature or non-existent, and provides energy storage researchers streamlined access to vital interphase-relevant data that can accelerate battery research efforts.

25 ENERGY STORAGE↗

Increasing Discovery and Usability of Earth Science Satellite Data with My NASA Data

For 20 years, the My NASA Data project at NASA Langley Research Center has developed innovative approaches to increase the use of NASA’s satellite data by learners. My NASA Data offers a variety of authentic Earth Science datasets and a data visualization tool, eliminating the need for educators and/or learners to obtain specialized knowledge of GIS data formats and software to access and use authentic Earth Science data. While there is no shortage of available data, as federal government agencies such as NASA house petabytes of freely accessible Earth Science datasets, much of the data are only available for download and visualization in specialized formats and software, limiting their accessibility to educators and learners, especially those in primary and secondary school. Using the Google Earth Engine platform, the My NASA Data team has recently reinvented their data visualization tool, called the Earth System Data Explorer (ESDE). The ESDE gives users the capability to explore over 60 Earth Science satellite datasets in a multitude of formats such as maps, graphs, and data table Its new and improved user interface design was developed based on the preferences of educators, whom the My NASA Data project has over 20 years’ experience working with. Earth Science and GIS Subject Matter Experts (SMEs) structured the data in a professional and scientific manner. During Fiscal Year 2023, the My NASA Data website received over 1 million digital engagements, with over one-third being visitors to the data visualization tool. These metrics highlight the interest in a visualization tool that is simple and free to use with reliable and trusted datasets. The ESDE empowers users to readily relate and analyze NASA Earth Science data within their area of interest. The team used a user-centered design (UCD) framework to receive and incorporate feedback into the application’s design. Core requested features include the ability to create time series graphs, comparative analysis of maps, and download the data as CSV file. Responses indicate that advances in data visualization tools such as the ESDE make authentic Earth Science data more accessible. This presentation will cover how the My NASA Data project develops tools to enhance data discovery and accessibility, as well as how SME and user suggestions are incorporated.

Desiray Wilson↗

FPGA Acceleration of GCN in Light of the Symmetry of Graph Adjacency Matrix

Graph Convolutional Neural Networks (GCNs) are widely used to process large-scale graph data. Different from deep neural networks (DNNs), GCNs are sparse, irregular, and unstructured, posing unique challenges to hardware acceleration with regular processing elements (PEs). In particular, the adjacency matrix of a GCN is extremely sparse, leading to frequent but irregular memory access, low spatial/temporal data locality and poor data reuse. Furthermore, a realistic graph usually consists of unstructured data (e.g., unbalanced distributions), creating significantly different processing times and imbalanced workload for each node in GCN acceleration. To overcome these challenges, we propose an end-to-end hardware-software co-design to accelerate GCNs on resource-constrained FPGAs with the features including: (1) A custom dataflow that leverages symmetry along the diagonal of the adjacency matrix to accelerate feature aggregation for undirected graphs. We utilize either the upper or the lower triangular matrix of the adjacency matrix to perform aggregation in GCN to improve data reuse. (2) Unified compute cores for both aggregation and transform phases, with full support to the symmetry-based dataflow. These cores can be dynamically reconfigured to the systolic mode for transformation or as individual accumulators for aggregation in GCN processing. (3) Preprocessing of the graph in software to rearrange the edges and features to match the custom dataflow. This step improves the regularity in memory access and data reuse in the aggregation phase. Moreover, we quantize the GCN precision from FP32 to INT8 to reduce the memory footprint without losing the inference accuracy. We implement our accelerator design in Intel Stratix10 MX FPGA board with HBM2, and demonstrate 1.3x-110.5x improvement in end-to-end GCN latency as compared to the state-of the-art FPGA implementations, on the graph datasets of Cora, Pubmed, Citeseer and Reddit.

Nair, Gopikrishnan R.↗