Graph algorithms are used in many big data and analytics workloads. The report presents performance using the PageRank algorithm. Oracle's SPARC M7 processor based systems provide better performance than an x86 E5 v3 based system.
Oracle's SPARC M7-8 server was able to deliver 3.2 times faster per chip performance than a two-chip x86 E5 v3 server running a PageRank algorithm implemented using Parallel Graph AnalytiX (PGX) from Oracle Labs on a medium sized graph.
The graph used for these results has 41,652,230 nodes and 1,468,365,182 edges using 22 GB of memory. All of the following results were run as part of this benchmark effort. Performance is a measure of processing rate, bigger is better.
8 x SPARC M7 (4.13 GHz, 8x 32core)
|281.1||3.2x faster per chip|
|x86 E5 v3 server
2 x Intel E5-2699 v3 (2.3 GHz, 2x 18core)
The number of cores are per processor.
Systems Under Test:
Graphs are a core part of many analytics workloads. They are very data intensive and stress computers. Each algorithm typically traverses the entire graph multiple times, while doing certain arithmetic operations during the traversal, it can perform (double/single precision) floating point operations.
The mathematics of PageRank are entirely general and apply to any graph or network in any domain. Thus, PageRank is now regularly used in bibliometrics, social and information network analysis, and for link prediction and recommendation. The PageRank algorithm counts the number and quality of links to a page to determine a rough estimate of the importance of the website.
Copyright 2015, Oracle and/or its affiliates. All rights reserved. Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective owners. Results as of October 25, 2015.