The supercomputer of a new type or a superconducting computer can be designed in the near decade. In appearance it will hardly differ from the currently operating models. But with its efficiency increased thousands of times the energy consumption will remain within reasonable limits by present standards. Besides, the sizes of a cooling facility will not increase either.

Specialists believe that efficiency of computers based on the existing transistor technologies has reached its utmost due to power limitations, so gigawatts of electric power will be required to start up a mechanism which is superior to modern analogs. It is comparable with the performance of a modern unit of a nuclear power plant and therefore is very expensive. The only alternative is a search for new approaches in modern electronics.

Application of superconducting facilities is one of the promising alternatives. Signals in superconducting electronics circuits can be transmitted practically with no losses, which favorably distinguishes this technology from transistor competitors. But to turn semiconductor-based computers from experimental prototypes to production samples they should be saved from an essential defect, i.e. low memory volumes. At present they do not exceed several kilobytes, while even a regular flash drive contains up to dozens and even hundreds of gigabytes. Specialists try to find a possibility to "construct" radically new memory cells for superconducting computers. In particular, by using magnetic materials which provide information storage for a long time and practically need no energy for this purpose.

A group of scientists from the Skobeltsyn Research Institute of Nuclear Physics (Moscow) has suggested application of the so-called hybrid method for "construction" of memory cells. Below is a description of the gist of the subject told by Igor Solovyov, Cand. Sc. (Phys. & Math.), senior research assistant of the nano-structure physics laboratory, to Yuri Drize, the Poisk newspaper correspondent, in his interview.

- "We are keen to exploit advantages of magnetic materials, which can register their status, and also merits of superconductivity-high, more than 100 GHz, characteristic frequencies and low, less than 1 microwatt, power consumption of logical circuits. The capacity of memory will increase but it will become very compact. Its volumes will compare with characteristics of the best modern computers or even will surpass them. Such memory can be relied on. Besides, the speed of computers will increase threefold in the short term and even sixfold in the long run."

According to Solovyov, the new computers will solve the same very intricate problems requiring processing of colossal data amount as the currently operating supercomputers, but on a higher level. For example, an

analysis of changes in the atmospheric behavior for long-term weather forecasting or genomic studies in biology. In medicine it can be modelling of propagation dynamics of different viruses and their mutations which took place over centuries. Superconducting computer will generalize information on evolution of serious illnesses such as tuberculosis or cancer for a period of time when physicians control them. And who knows maybe the data processed by the computer will suggest new treatment modes.

How is it assumed to achieve singular properties of the new device?

- Solovyov noted: "We are obliged to a possibility to work with the competitive phenomena such as magnetism and superconductivity to a breakthrough in this field. It was not possible earlier as magnetic materials actually killed superconductivity. But in recent years we managed to create ultrafine magnetic films which do not destroy superconductor effects but only influence them in a specific way. An unusual effect is achieved by combining of superconducting films and magnetic material films in one nanostructure. In particular, memory capacity can increase millionfold. Such are potentials of the most advanced technology to date.

Its origin is linked with the works of a scientific team of the Institute of Solid State Physics (Chernogolovka, Moscow Region) headed by Valery Ryazanov, Dr. Sc. (Phys. & Math.). In 2001, for the first time in the world they produced a hybrid nanostructure consisting of superconductors and ferromagnetics, and this fact was disclosed by the Nature journal. But the sense of understanding of how to use the achievement came later. Besides, those years were rather hard for science in our country. The new technology was fast developing in the USA and Japan, first of all due to government financing programs. But in Russia the research work was carried out actually on the personal initiative of the developers, and therefore the experiments were unavailable. This problem was handled, apart from the two above said institutes, by specialists headed by Lenar Tagirov, Dr. Sc. (Phys. & Math.), at the Kazan (Volga Region) Federal University which received excellent technological equipment. Today the situation changes for the better, i.e. funds are appropriated for research and experiments. But in contradistinction, for example, to the USA there is no special state financing program in our country for these works intended for several years.

This problem was discussed at the seminar "New Solutions for Energy-Effective High-Performance Computing Systems" held by the Foundation for Perspective Studies in August of 2014. The report on the development of superconducting electronics for receiving-trans-

forming and computing systems was made by Director of the Lukin Research Institute of Physical Problems Alexander Gudkov, Cand. Sc. (Phys. & Math.). The seminar resolution stated that from 2015 the Foundation would start implementation of a project for creation of a Russian ultrahigh-performance supercomputer working on new physical principles.

Solovyov stressed: "Head of our laboratory Mikhail Kupriyanov, Dr. Sc. (Phys. & Math.), has been engaged in studies of hybrid structures (superconducting and magnetic materials) for many years, and he is an ideological inspirer of our research team. In cooperation with the associate professor of the physics department of Moscow State University Nikolai Klyonov, Cand. Sc. (Phys. & Math.), we are carrying on joint theoretical studies for about seven years. Only a solid theoretical basis will provide a real advance. In particular, it concerns experimental production of a basic component of perspective memory elements, i.e. composite thin-film nanostruc-ture consisting of several layers of different metals and dielectrics. Its thickness makes several nanometers. It is clear that its handling needs only special devices including an electronic microscope, which offers an opportunity of analyzing material at an atomic scale.

It is true that the production difficulties prevent yet from development of very complicated "sandwiches" consisting of almost a dozen of ultra-fine layers as it was managed to achieve by specialists from one of the American companies. But by means of a careful theoretical analysis and properly selected materials of a simpler (five-layer) structure our scientists plan to produce elements exceeding the world analogs by fast operation. Besides, if they succeed in achieving error "resistance", it will be possible to create a superconducting energy-efficient compact magnetic memory with random access. This problem is being tackled now by the laboratory workers.

Composite materials will find also other applications. Traditionally the sophisticated sensors are based on superconducting technologies. They are used in development of tomographs of a new generation, microscopes, magnetometers and other precision instruments necessary for studies of extra-fine materials. This technology will thrust forward research works in different scientific spheres, primarily in nanotechnologies and also in medicine and biology.

Yu. Drize, Will Keep in Mind. Potentialities of Supercomputers Will Increase by Orders of Magnitude.- "Poisk" newspaper, Nos. 33-34, 2014

Photo supplied by Igor Solovyov

Prepared by Sergei MAKAROV