13 Jul 2011 Moscow - Russia's most powerful supercomputer, Lomonosov, installed at Moscow State University, outperformed all competitors during the recent Graph500 benchmark tests. Graph500 was initiated by experts from a number of national research centres, acting under the auspices of Sandia National laboratories (USA). The ratings are compiled on the basis of tests carried out on high-performance computer systems. The goal of this particular benchmark is to determine which supercomputers are most efficient at processing arrays of sparse data, presented either in the form of a graph or as a database.
Such tasks are commonplace in pharmacology, information security, and many other knowledge-driven industries. In order for these complex tasks to be resolved effectively, high-performance systems capable of processing hundreds of petabytes of data are required. In the course of the tests, the time spent by each system to process the graph is assessed. This result is measured in terms of the number of edges traversed in a unit of time. When the test results were analyzed, it was found that the Lomonosov, a Russian supercomputer designed by T-Platforms, had set a new performance record, reaching a level of 43.5 GE/s (billion edges processed per second).
"The Graph500 ratings are interesting because they allow us to assess not only the computing power of supercomputers but also their efficiency, in real terms, at processing huge quantities of data. In such tests, those computer systems that have been optimized for working with large arrays of data will always triumph", stated Vsevolod Opanasenko, CEO of T-Platforms. "The Lomonosov supercomputer has proven its efficiency in a number of scientific projects carried out at Lomonosov Moscow State University and shown outstanding performance during the latest Graph500 Benchmark tests. This is yet another endorsement of Russia's achievements in creating high-performance systems aimed at solving an extremely wide range of applied problems."
"Emerging applications in health care informatics, social network analysis, web science, and anomaly detection in financial transactions, require a new breed of data-intensive supercomputers that can make sense of massive amounts of information", stated David A. Bader, Professor at the Georgia Institute of Technology, one of the founders of the Graph500 rating. "The Graph500 Benchmark highlights the importance of new systems that can find the proverbial needle in the haystack of data. The second list recently released at ISC'11 contains 28 entries from the United States, Europe, China and Russia. IBM's BlueGene/P placed first by solving the largest graph with over 250 billion vertices, yet received strong competition from the Russian T-Platforms supercomputer in Moscow that demonstrated the highest aggregate performance of any system at 43.5 billion edges of the graph traversed per second. With these capabilities, a machine on the top of this list may analyze huge quantities of data to provide better and more personalized health care decisions, improve weather and climate prediction, and better integrate our on-line social networks with our personal lives."