Composing Data Parallel Code for a SPARQL Graph Engine

Castellana, Vito G.; Tumeo, Antonino; Villa, Oreste; Haglin, David J.; Feo, John

doi:10.1109/socialcom.2013.104

Home/Archive/DOE OSTI

DOE OSTI · 1123246

Composing Data Parallel Code for a SPARQL Graph Engine

Castellana, Vito G.·Tumeo, Antonino·Villa, Oreste·Haglin, David J.·Feo, John

View original source ↗Export metadata ↓

Abstract

Big data analytics process large amount of data to extract knowledge from them. Semantic databases are big data applications that adopt the Resource Description Framework (RDF) to structure metadata through a graph-based representation. The graph based representation provides several benefits, such as the possibility to perform in memory processing with large amounts of parallelism. SPARQL is a language used to perform queries on RDF-structured data through graph matching. In this paper we present a tool that automatically translates SPARQL queries to parallel graph crawling and graph matching operations. The tool also supports complex SPARQL constructs, which requires more than basic graph matching for their implementation. The tool generates parallel code annotated with OpenMP pragmas for x86 Shared-memory Multiprocessors (SMPs). With respect to commercial database systems such as Virtuoso, our approach reduces memory occupation due to join operations and provides higher performance. We show the scaling of the automatically generated graph-matching code on a 48-core SMP.

Explore related subjects

Big data SPARQL SPARQL-to-C

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Castellana, Vito G., Tumeo, Antonino, Villa, Oreste, Haglin, David J., Feo, John. 2013-09-08. Composing Data Parallel Code for a SPARQL Graph Engine. https://doi.org/10.1109/socialcom.2013.104

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

DOE OSTI2022

Machine Learning Using a Simple Feature for Detecting Multiple Types of Events From PMU Data

This paper describes simple and efficient machine learning (ML) methods for efficiently detecting multiple types of power system events captured by PMUs scarcely placed in a large power grid. It uses a single feature from each PMU based on a rectangle area enclosing the event in a given data window. This single feature is sufficient to enable commonly used ML models to detect different types of events quickly and accurately. The feature is used by five ML models on four different data-window sizes. The results indicated a tradeoff between the execution speed and detection accuracy in variety of data-window size choices. Here, the proposed method is insensitive to most data quality issues typical for data from field PMUs, and thus it does not require major data cleansing efforts prior to feature extraction.

Big data↗

DOE OSTI2015

Open Research Challenges with Big Data - A Data-Scientist s Perspective

In this paper, we discuss data-driven discovery challenges of the Big Data era. We observe that recent innovations in being able to collect, access, organize, integrate, and query massive amounts of data from a wide variety of data sources have brought statistical data mining and machine learning under more scrutiny and evaluation for gleaning insights from the data than ever before. In that context, we pose and debate the question - Are data mining algorithms scaling with the ability to store and compute? If yes, how? If not, why not? We survey recent developments in the state-of-the-art to discuss emerging and outstanding challenges in the design and implementation of machine learning algorithms at scale. We leverage experience from real-world Big Data knowledge discovery projects across domains of national security, healthcare and manufacturing to suggest our efforts be focused along the following axes: (i) the data science challenge - designing scalable and flexible computational architectures for machine learning (beyond just data-retrieval); (ii) the science of data challenge the ability to understand characteristics of data before applying machine learning algorithms and tools; and (iii) the scalable predictive functions challenge the ability to construct, learn and infer with increasing sample size, dimensionality, and categories of labels. We conclude with a discussion of opportunities and directions for future research.

Big data↗

DOE OSTI2015

Using Clinical Data, Hypothesis Generation Tools and PubMed Trends to Discover the Association between Diabetic Retinopathy and Antihypertensive Drugs

Diabetic retinopathy (DR) is a leading cause of blindness and common complication of diabetes. Many diabetic patients take antihypertensive drugs to prevent cardiovascular problems, but these drugs may have unintended consequences on eyesight. Six common classes of antihypertensive drug are angiotensin converting enzyme (ACE) inhibitors, alpha blockers, angiotensin receptor blockers (ARBs), -blockers, calcium channel blockers, and diuretics. Analysis of medical history data might indicate which of these drugs provide safe blood pressure control, and a literature review is often used to guide such analyses. Beyond manual reading of relevant publications, we sought to identify quantitative trends in literature from the biomedical database PubMed to compare with quantitative trends in the clinical data. By recording and analyzing PubMed search results, we found wide variation in the prevalence of each antihypertensive drug in DR literature. Drug classes developed more recently such as ACE inhibitors and ARBs were most prevalent. We also identified instances of change-over-time in publication patterns. We then compared these literature trends to a dataset of 500 diabetic patients from the UT Hamilton Eye Institute. Data for each patient included class of antihypertensive drug, presence and severity of DR. Graphical comparison revealed that older drug classes such as diuretics, calcium channel blockers, and -blockers were much more prevalent in the clinical data than in the DR and antihypertensive literature. Finally, quantitative analysis of the dataset revealed that patients taking -blockers were statistically more likely to have DR than patients taking other medications, controlling for presence of hypertension and year of diabetes onset. This finding was concerning given the prevalence of -blockers in the clinical data. We determined that clinical use of -blockers should be minimized in diabetic patients to prevent retinal damage.

Big data↗