by Marina KHALIZEVA, journalist

A 420 teraflop supercomputer was commissioned at Moscow State University in November 2009. The largest in Eastern Europe, this engine is capable of handling as many as a trillion (thousand billions) operations per second. It ranks twelfth in the Top-500 list of the world's best computer systems.

This is the nation's first venture of a class like that. It was realized by the Moscow-based T-Platforms Company* leading in the Russian market of high-performance technology. The new supercomputer, a record-beater in many areas, incorporates three types of nodes: T-Blade 2, T-Blade 1.1, and PowerXCell. The first platform developed by our engineers virtually from scratch accounts for 90 percent of the machine's effi-ciency. The company designed and patented the essential boards and mechanical components. According to Vsevolod Opanasenko, T-Platforms Director General, this hardware, which is in the highest efficiency range (1.1 petaflop**), consolidates our country's hold in the world's supercomputerland. Launched at Moscow University, the computer that takes as much as 252 m² floor-space (!) is just 4 times as slow as the world's greatest Gray XT Jaguar in the USA (1.75 petaflop). It was quite natural for Lomonosov Moscow State University to name its superengine after Mikhail Lomonosov, the great Russian savant and scientist of the 18th century.

The capacity of its core (main) and disk memory is 56,676 gigabytes and 166.4 terabytes respectively, and its storage capacity, equal to 350 terabytes.

Russia is an active player on the hard- and software market. While in June 2007 she had but three superpow-er machines, in November the same year she had as many as seven, then eight; today this country has upped this number to fifteen. The greater part of this machin-ery has been set up with state support at major universi-ties, research centers and defense departments. These electronic computers are becoming an indispensable tool in many research areas such as hydraulic and gas dynamics, biological systems and weather-wherever

* See: S. Abramov, "Supercomputers of a New Generation", Science in Russia, No. 3, 2009.-Ed.

** Flop (flops)-Floating Point Operations Per Second. This value, used for measuring a computer's efficiency shows how many floating point operations it can carry out per second; petaflop = 10" flops; tera= 10¹².-Ed.

one has to compare a large set of interdependent and time-variable parameters. Automanufacturers, aircraft builders and oil-prospecting companies need such com-puter know-how. The garment industry and fashion designers want it, too, to keep their market going. And if we take nanotechnologies, they cannot do without supercomputers either: many processes, not amenable to measurement, could be modeled only virtually. Supercomputers are in demand in the banking and financial sectors (transactions, risk factors), telecom-munications (cryptography, internet infonautics, tele-phone).

Moscow University scientists working for strategic branches of the national economy like power and ener-gy saving, medicine, the drug industry, aeronautics and space put most of the computing job on the "iron shoul-ders" of computers. In March 2008 the university's computing center put into service a unique setup named CHEBYSHEV-in honor of the great Russian mathe-matician of the 19th century (the engine was assembled by the RAS Institute of Program Development Systems and T-Platforms Company); it can handle as many as 60 trillion (thousand billions) operations a second. Soon after, IBM (USA) commissioned the BLUE GENE, its capacity being 28 teraflops. Both giants, working full tilt, complement each other. While CHEBYSHEV is geared to basic research in bio- and nanotechnologies, medicine, simulation of new materials and brain research, BLUE GENE is hooked to scientific research and education. College students have a chance to learn the ropes of work at such systems and upgrade their pro-gramming skills.

CHEBYSHEV has helped speed up work on a new thrombolytic, a drug that prevents thrombosis (clot for-mation) in blood vessels: this work, completed in less than a year and a half, would have taken a decade with the use of conventional methods. The new medicine went through clinical tests in 2008. Quite recently this

superpower computer reran the scenario of the Mount Vesuvius eruption in southern Italy that buried the ancient Roman city of Pompeii in A.D. 79. The giant engine was the first to demonstrate how it happened: at first sea water flooded the hot volcanic magma and then, five hours later, the second eruption of the volcano buried the city.

The LOMONOSOV computer complex (that escalat-ed Moscow State University to the club of the world's largest supercomputer centers) will be cracking harder problems than that. Basic medicine doctors researching in the individual variability of the genome and pro-teome* and in biomarkers** for the diagnostics and therapy of socially significant diseases will get a welcome aid in LOMONOSOV. The point is that these studies generate a plethora of experimental data snowballing exponentially. It is very difficult to digest this ocean of information-read all the articles on the subject, for instance. This is where high-performance computer sys-tems come in handy to take care of data evaluation and management towards practical uses in biomedicine and healthcare. They cater to a felt need. In fact, it will be just impossible to cope with the job at hand without LOMONOSOV, a superengine capable of storing signif-icant amounts of information, and working very fast.

It was no easy job to build a supercomputer like that into the available infrastructure. It's like putting up a real industrial enterprise. As Acad. Viktor Sadovnichy, Moscow University Rector, says, the University has concentrated immense material and intellectual resources on this project. Even the present economic stalemate, when state subsidies had to be trimmed, did not arrest it: the University shelled out for essential con-struction and repair works on three stories of the build-ing and completed the assembly of the essential techni-cal systems. As much as 5 MW power is supplied to keep the 420 teraflop computer at work.

Moscow State University, the Number One educa-tional institution in Russia, shoulders great responsibil-ity for the further progress of computer technologies. It is training computer experts at four academic depart-ments-that of Mechanics and Mathematics, Computer Mathematics and Cybernetics, Physics, and Bioinfor-matics and Bioengineering. Having supertechnology in the 0.5 to 420 teraflop range, Moscow University is arranging practicals within its walls. Colleges of further education to be opened soon in Nizhni Novgorod, Tomsk, Chelyabinsk and other major university cities of the nation will be tied into basic departments running M. Sc. courses and practicals, and supplied with ade-quate literature, teaching aids and internet media. It's a fact: Russia is moving into a new, higher level of com-puter power.

* See: A. Archakov, "The Time of Proteomics", Science in Russia, No. 3,

2009.-Ed.

** See: V. Tkachuk, "Medicine of the Future", Science in Russia, No. 3,

2008.-Ed.