A supercomputer is a computing system able to manage vast data and perform highly complex tasks with extreme speed and precision. Such machines consist of many processors working together in parallel, linked through a single network. Supercomputers find use in scientific research, simulations, encryption, artificial intelligence, and other areas that demand immense computing power.
On September 1, Moscow State University named after MV. Lomonosov unveiled and commissioned a new unique machine called MGU-270. This system will support scientific research in artificial intelligence. It houses nearly a hundred latest generation graphics accelerators, delivering rapid and efficient computation. Rector Viktor Sadovnichy stated that no other university possesses equipment of this scale.
In an interview with socialbites.ca, Vasily Fomichev, deputy dean for scientific work at the VMK faculty, confirmed that the supercomputer is located at the Faculty of Computational Mathematics and Cybernetics (VMC).
“The system is ready, tested, and practically operational. There is a large machine room with shelves, plus engineering infrastructure, power supply networks, and advanced cooling because a supercomputer consumes substantial energy,” explained the scientist.
Fomichev noted that the project involves more than just the main calculating unit. There is a significant supporting setup including dispatcher rooms, high-power electrical facilities, uninterruptible power supplies, and battery rooms.
supercomputer power
In the days leading up to the launch, several outlets quoted the university rector about the machine being among the world’s most productive. “On September 1 we officially open a new supercomputer at Moscow University built by our team. It will reach hundreds of petaflops in capability and may rank second or third globally in power,” the rector stated. These comments came through TASS.
Yet Fomichev urged caution, saying the publicized numbers need careful interpretation. “Reports mentioned 400 petaflops of capacity and a top global position, but those figures are not directly comparable with standard benchmarks. They reflect design capability rather than measured performance,” he explained.
One common way to gauge a supercomputer’s performance is the LINPACK benchmark, which tests how quickly a system can solve large systems of linear equations. The outcomes are published in the TOP-500 list, updated twice a year.
According to the scientist, the declared 400 petaflops relates to design intent rather than a confirmed LINPACK result. The practical speed under real testing would be different.
What does the MGU-270 do?
The MGU-270 was developed by several divisions of Moscow State University under the leadership of VMK faculty experts. It aims to tackle complex scientific and socio-economic challenges and to advance basic and applied research in artificial intelligence.
MSU already operates systems such as Lomonosov and Lomonosov-2, which address mathematical modeling of physical phenomena. These models simulate liquid and gas flows, turbulent processes, and the behavior of substances across states.
The team notes that classical computer architectures remain essential for simulating real physical events, but many modern problems involve big data, imagery, text, and data streams. New architectures optimized for AI, neural networks, and large data sets are needed. The MGU-270 fits into this shift and is part of a broader family of powerful university machines. Fomichev also remarked that universities nationwide still do not have comparable classroom-grade systems.
Fomichev added that the MGU-270 will also support modeling physical phenomena, but its primary focus will be on information analysis and recognition tasks.
For example, providing the MGU-270 with large MRI and CT image datasets could enable the system to assist in diagnosis by learning patterns in the images. The teams emphasize teaching computing systems to address diagnostic challenges quickly while relying on large data archives and the accumulated experience of researchers. The aim is for the new supercomputer to help accelerate such capabilities.
Officials from TASS noted that the machine is already tackling image analysis and medical testing tasks, with initial results expected in December.
Another promising line of work involves developing tools to gauge student attention in classrooms. Researchers hope to create systems that analyze behavior and alert teachers when attention dips. Beyond education, the project is expected to speed up material development, space exploration, medical advances, and other fields that require substantial computational resources.