Lucata Pathfinder Demonstrates Breakthrough Graph Analytics Processing Efficiency by Ranking #46 on Green Graph500 Benchmark
Ranks #211 on Graph 500 BFS Benchmark
New York City – November 18, 2021 – Lucata, provider of a next generation computing architecture based on Intel technology for high performance, massively scalable processing of graph analytics, today announced that Lucata Pathfinder ranked #46 on the Green Graph500, which ranks graph analytics processing performance per energy consumed (GTEPS/Watts). Lucata Pathfinder ranked #211 on the Graph 500 BFS Benchmark, which measures raw graph analytics processing performance, delivering nearly 1 GTEP (845 MTEPS) at scale 30. The Georgia Tech Center for Research into Novel Computing Hierarchies (CRNCH) ran the benchmark testing using four Lucata Pathfinder chassis.
The Graph 500 BFS benchmarks test the performance of graph analytics processing systems on synthetic Breadth-First Search (BFS) datasets, enabling comparisons of a broad range of optimized hardware/software solutions for graph analytics processing. The benchmarks help in the design and procurement of hardware architectures and software systems for data-intensive graph analytics processing on large, sparse matrix datasets.
With energy consumption rapidly becoming a limiting factor in the future of big data computing, an energy-aware benchmark has also become a necessity. The Green Graph500 provides performance-per-watt metrics for the Graph 500 participants. It acts as a forum for vendors and data center operators to compare the energy consumption of data-intensive graph analytics computing workloads on various architectures. Energy consumption is a crucial metric that drives operating costs for large scale, graph processing systems.
“Next generation computing solutions like Lucata Pathfinder can demonstrate a quantum leap in performance by submitting benchmark results to the vendor-agnostic Graph 500. The benchmarks allow vendors such as Lucata to improve our products while enabling buyers to make fair comparisons between graph analytics processing solutions,” said Lucata COO and Chief Technologist Marty Deneroff. “These new Graph 500 results demonstrate the massive performance improvements Pathfinder is able to deliver. Pathfinder requires 1/8th the rack space and consumes 1/10th the power of conventional computing approaches. We are especially proud of the Green Graph500 results, which demonstrate the enormous energy consumption benefits of the Lucata architecture.”
Georgia Tech purchased two Lucata Pathfinder chassis in support of their National Science Foundation (NSF) CISE Community Research Infrastructure (CCRI) project, “A Community Research Infrastructure for Post-Moore Computing.” The research team for this project, which is led by Senior Research Scientist Dr. Jeffrey Young, is using the Pathfinder chassis to continue its research into next generation computing architectures that can overcome data access challenges that occur for sparse Big Data workloads. Access to the two Georgia Tech chassis is available to U.S.-based researchers via an application on the CRNCH webpage. Lucata loaned Georgia Tech two additional Pathfinder chassis for the Graph 500 testing and to help further next generation computing research in general at CRNCH.
“We are very excited to see the results of graph analytics benchmark testing and the relative performance capabilities of Lucata’s next generation Pathfinder architecture,” said Dr. Young. “As we look to solve large, sparse problems like those represented by Graph 500, we are interested to maximize the benefits of novel “post-Moore” architectures like Lucata’s migratory thread system. We look forward to further enabling our students and other research organizations to fully explore the capabilities of the Pathfinder system for a wide array of potential use cases.”
Pathfinder enables organizations to leverage massive pools of physical memory to accelerate and scale graph analytics and AI and ML model training by orders of magnitude beyond the capabilities of conventional computing approaches. The solution enables high-performance exascale graph analytics, including exhaustive breadth-first search (BFS), on unpruned, unsharded massive graph databases. Lucata can be used with open source or commercial graph software or with custom-written graph solutions that leverage LAGraph, GraphBLAS, or the Lucata library of search algorithms, enabling organizations to use their existing software to uncover much deeper connections within much larger graphs than possible today. These unique capabilities allow organizations to reimagine the potential of graph analytics, AI and ML and address intractable challenges in fraud detection, cybersecurity, blockchain, risk assessment, healthcare and many other fields. Ultimately, the Pathfinder hardware architecture will be migrated to an ASIC design, which will deliver a 10x increase in processing performance, enabling extreme high performance for a broad range of common Big Data computing use cases which cannot be cost-effectively addressed with conventional computer architectures.
More information about the next-generation Lucata computing platform is available on the Lucata website including:
- Lucata Pathfinder description
- Lucata Pathfinder datasheet
- Lucata Tech Brief on next generation computing
- Lucata performance benchmarks
- Lucata use cases
The Lucata next generation computing architecture leverages Intel technology to enable organizations to accelerate and scale graph analytics orders of magnitude beyond the capabilities of conventional computing approaches. Lucata leverages patented Migrating Thread technology to massively scale unified memory and conduct high-performance graph analytics, including exhaustive breadth-first search (BFS), on massive unpruned and unsharded databases. Organizations can now use their existing graph database software or custom graph solutions to analyze deeper connections on much larger graphs than ever before possible using conventional servers. Lucata solutions support groundbreaking graph analytics and improved machine learning and AI for organizations in financial services, cybersecurity, logistics, blockchain, healthcare, life sciences research, telecommunications, ecommerce, government and more. The company has offices in Palo Alto, New York City, and South Bend, Ind.
Sr. Product Manager, Lucata