MOSCOW, April 6 -- The 12th edition of Russia's Top50 list of the most powerful HPC systems was released at the annual "Parallel computing technologies" international conference last week. The new list is topped by "Lomonosov" supercomputer with peak performance of 414 teraflops. The system deployed at Moscow State University (MSU) occupies the 12th position on the current TOP500 list of the world's most powerful supercomputers.

On March 30, 2010, the Research Computing Center of Lomonosov Moscow State University and the RAS Joint Supercomputer Center have released the 12th edition of the Top50 list of the most powerful HPC systems of Russia and the CIS. The list was announced at the annual "Parallel computing technologies" international conference held by the Supercomputing Consortium of Russian Universities and the Russian Academy of Sciences under patronage of the Russian Foundation for Basic Research.

The new rating is topped by "Lomonosov" supercomputer manufactured by T-Platforms, the leading Russian HPC vendor. The system with peak performance of 414 teraflops and Linpack-measured performance of 350 teraflops occupies the 12th position of the current TOP500 list of the world's most powerful computers. The system deployed at a modest footprint of only 252 square meters is among the densest HPC installations in the world, thanks to the innovative T-Blade2 platform capable of achieving impressive 18TFlops per rack with Intel Nehalem processors.

Over 250 research groups at MSU employ Lomonosov for basic and applied science solving computational problems of magnetic hydrodynamics, hydro- and aerodynamics, quantum chemistry, seismic processing, geology, nanoscience, and cryptography. Among the first jobs submitted for computation were problems of sound generation by turbulence and computer-aided drug design.

Lomonosov has knocked the previous Top50 leader, MVS-1000K supercomputer installed at the RAS Joint Supercomputer Center with 107 teraflops of Lipack performance, off the top perch to the 2nd position. The 3rd position of the new list is occupied by SKIF Chebyshev 60 teraflops system developed by T-Platforms and deployed at MSU in the framework of the SKIF GRID supercomputer program of Russia and Belarus.

For the first time history the bottom performance level of the list exceeded 1 teraflops to reach 1.47 teraflops. The number of systems used in science and education decreased slightly 31 to 28, while one new system involved in applied research has joined the list. The number of systems employed for financial analysis grew from 3 to 5, while the industrial manufacturing segment lost one system (down from 6 to 5) and the number of systems involved in applied research has grown (from 9 to 10).

T-Platforms supplied 30 percent of systems on the list (15), followed by HP (13 installations) and IBM (11 systems). T-Platforms leadership however is much more prominent with regard to the overall system performance: the share of systems supplied by T-Platforms exceeds 50 percent of the overall 888 teraflops Linpack performance of the list.

The next edition of the Top50 list of the most powerful supercomputers of Russia&CIS will be announced at the end of September 2010 at the international scientific conference "Scientific service on the Internet: supercomputer sites and problems."

-----

Source: T-Platforms