Site: Scientific Research Institute
of Computer Complexes (NIIVK)
108 Profsoyuznaya St.
Telephone: 095 330-0992-7401
Fax: 095 434-0065
Date Visited: May 11, 1993
Report Author: D. Blidberg
Alexander Krupsky;Director for Research and Development
Juri Mordovski; Director, Sigma Technologies
Telephone: 095 924-1654, 921-6396
Fax: 095 924-1654
The meeting at the Scientific Research Institute of Computer Complexes (Nauchno-Issledovatesky Institute Vychislitelnyh Kompleksov or NIIVK) was held with representatives of NIIVK and other companies that were using the software that had been developed at the institute. An overview of the institute and its activities was presented and then various applications were described by members of the companies that had developed or were developing those applications. The institute was represented as the organization in Russia that was responsible for designing all computer hardware and software. It did not build many of the subsystems, but participated in the design, development, and system integration. It seemed that the institute was guided by the needs of radar and sonar processing as well as the missile warning system. Recently some of those applications (such as the missile warning system) were unclassified, but many of the sonar and radar processing applications remain classified.
NIIVK was founded in the 1950s to develop computer hardware and software systems. In 1976 the institute moved to its current location. NIIVK has developed four to five generations of large computers (supercomputers). Originally the institute focused on developing computers for classified applications, but recently it has focused on some unclassified applications. NIIVK's scientists developed the early warning system for missile attack that is now unclassified, but continues to develop some systems for radar and sonar applications that remain classified. The following summary will give some idea of the focus of the institute.
The Scientific Research Institute of Computer Complexes, situated in Moscow, is a member of the Scientific Industrial Association for Computers (NPO SVT), the largest such association in Russia.
The staff of NIIVK was formed in the early 1950s. After creating the M-2 and M-4 computers (two of the first Soviet computers), the serial semiconductor computer M-4-2M was designed in 1964 and was serially produced until the end of the 1980s. The computer M-10, designed by the NIIVK in 1975, contained parallel architecture. Novel technical ideas were realized in this computer that allowed noted scientist and computer architect M.A. Kartsev, the first director of NIIVK, to achieve a high performance. American researchers P. Wolcott and S. Goodman wrote: "The N-10 is distinguished by its ability to conveniently parallel-process data of different formats, dynamically changing the clustering of processors to match the format of the data at hand."
The newspaper ComputerWorld USSR gave the following estimation of the computer M-10: "Apotheosis of heroic period of the Soviet informatics was a creation of the powerful 64-bits vector-pipeline supercomputer M-10...."
The original trends of the NIIVK supercomputer are characterized by usage of so-called "type M" architecture, the first representative of which was computer M-10. This architecture is notable for using a wide-format memory that exchanges vector operands with several identical processors, executing the same instruction stream. A later stage in the development of the type M computer architecture was the creation of the new vector multiprocessor supercomputer M-13, which was produced serially beginning in 1985.
The peculiarities and architectural principles of computers developed by NIIVK provide high performance for real time data flow processing with large amounts of information (real time radar and sonar signal processing, image recognition, etc.). These peculiarities and principles are as follows:
Architectural principles developed and realized by NIIVK in the M-10 system were used in the Burroughs BSP Computer. Some technical decisions in vector processing coincide with those in Cray and CONVEX computers. Also, some memory organization features are similar to capability-based systems (like Plessey System 250 and Intel iAPX 432).
NIIVK has designed software products for its computers, including the following:
The different models of computers designed by NIIVK have been produced serially for nearly thirty years and work with high reliability, providing all requirements of the most crucial networks and systems up to the present.
Because of the need to model sound propagation fields, a new computer design has been undertaken. The new machine is referred to as a database machine. The system design was completed and construction had begun; however, construction has been stopped due to lack of funding. The machine deals with numeric data, textual (or symbolic) data, and other information. This new computer was felt to be the system on which artificial intelligence (AI) systems might be implemented as well as other software. AI may be used for identification and classification applications in the future.
When asked about the data to be included in the database machine, the team was told that the institute focuses only on the design of hardware and software systems. However, the institute's scientists have collected various oceanographic data from the Atlantic Ocean that has been included in the new computer design efforts.
Mr. Juri Mordovski, Director of Sigma Technologies, discussed work underway at the company. Sigma Technologies was started two years ago to focus on the development of AI and the transfer of that technology to the United States and other markets. The company undertook several applications, including two that were suggested by the Director of NIIVK, Mr. Alexeev. Some of the applications suggested were the development of a system for monitoring large mechanical systems, and development of AI systems for automatic monitoring of other complex systems, and automatic surveying. When Mr. Mordovski was asked if he knew of any AI efforts being undertaken in Russia related to undersea systems, he stated that as far as he knew, "there is no work in Russia where AI is being applied to undersea systems." His company had recently undertaken a summary of AI work underway in Russia and found only work in expert systems.
The company has developed some automatic translation tools that are now available. One discussed was a translator to translate legal documents from Russian to English in real time over electronic mail (E-mail). The company has been working on this with Lockheed and several other U.S. companies.
Mr. Anekyn is the head of a small group in Volna that is focusing on the development of sonar systems utilizing some of the work produced at NIIVK.
Algorithms with super resolution have been developed in the institute. This high resolution is possible because the institute's scientists have been able to eliminate all surface and bottom reverberations, greatly suppressing unwanted returns.
NIIVK has databases of oceanographic data; physical and geological data; bottom and surface data; and hydrological water parameters, such as CTD information.
Bottom and surface reverberation is understood and mapped in this system.
Development of medical imaging systems is underway at the Cardiological Center of the Russian Academy of Sciences. The center has a 5 MHz sonar that develops 2-D images. Scientists can combine those 2-D slices to develop 3-D images now, but hope to develop a system that will produce 3-D images in real time.
As a result of visits to the United States, specifically NASA, NIIVK representatives assess their capability as being up to date in the area of software development but not in terms of hardware development. Their computers seem to be less capable than comparable U.S. hardware, but the Russians make up for that disadvantage by increased sophistication of microprogramming and more complex control devices within the computer hardware. There seems to be a concerted effort to bring some of the developed algorithms into commercial products. These, when developed, will be very capable systems.
The NIIVK representatives with whom the WTEC team met were very helpful and encouraged further cooperation in all of the areas discussed. They want to find methods by which to disseminate information about their work, such as professional publications to which they might submit their results. There seems to be much good work to better understand. Time limitations constrained in-depth discussions of their various activities.