Mikhail A.Kartsev - Developer of Super-Computers for Space Observation

 

Mikhail Kartsev belong to the kind of scientists who, for some reasons, acquired official and full acknowledgment of their large achievements only after their death. The academic elite did not award him any special titles. It wasn't until ten years after his death that the institute that he himself founded, Moscow's Scientific Research Institute of Computer Complexes (SRICC), was renamed after him.

Computer science and technology were really his mission. It brought him both creative satisfaction and grief. He devoted it all his time - whill working and resting.

Mikhail Kartsev was born into a teacher's family in Kyiv on May 10, 1923. His father died the same year. Michael lived with his mother in Odessa, Kharkov, then from 1934 to 1941 - in Kyiv again, where he finished secondary school. In summer 1941 he was engaged in fortification defense work in Donbass. In September he was called to the army, where he served until February 1947. During the Great Patriotic War, tankman Kartsev fought in the South-West, South, North Caucasian and on two Ukrainian Fronts. He participated in the liberation of Romania, Hungary, Czechoslovakia and Austria. For courage demonstrated in battle, in his twenties, he was awarded the medal "For Bravery", the Red Star Order, a medal "For the Conquest of Budapest" and "For Victory over Germany".

After demobilization, Kartsev entered Moscow Power Enginiring Institute (MEI) the Radio-technological Department. In his third year of study he passed examinations without attending lectures, one year ahead of schedule and in 1950, being a 5th year student, he started work (on a part-time basis) at the Laboratory of Electrical Systems of the Power Engineering Institute of the AS of USSR, where he participated in development of one of the first Soviet computers - the M-1. In 1952 he was appointed to the permanent position of junior scientific assistant at the same lab of the Power Engineering Institute of AS USSR. Working on the design of the M-2 computer, he demonstrated extraordinary capability. And, the computer was created by a small team in only one and half years! (The BESM computer took twice as long time and was developed by a much more powerful team). Although the M-2 was not as powerful as BESM, Kartsev himself described it as a "solid computer".

In 1957 the development of one of the first in the USSR transistor computer started.

In November 1962, a government order on M-4 the mass-manufacturing was issued. However, Kartsev supported by his team, proposed to postpone putting a new computer for mass-manufacturing. He offered to eliminate the 'bugs' of the agreed-upon model, hoping to make it more technically efficient during its production and adjustment. Besides, at that moment, they had just developed a new system of logic elements using high frequency transistors which could help the units operate much more rapidly, and then, there also appeared power transistors which could cancel the usage of radio tubes altogether.

The argument was terminated by a Decision of the Military Industry Commission at the Presidium of the Cabinet Ministers of the USSR, issued in March 1963. And, in the same month we prepared and passed documentation for the computer's first frame (the arithmetic unit) to a factory now under V.Kurochkin's management. These computers were manufactured for a period more than 15 years and probably are still under operation.

Ahead of His Time

It seemed that the time same to calm down and take a rest, or at least, to have a small break after all the hard work.

However, this was not done and probably could not have been done. In 1966, Kartsev had put forward an idea of a multi-computer complex still consisting of computers specially designed for such joint work. Preliminary conducted research demonstrated that the computer might achieve an operational speed of one billion op/s. At the time, there was no computer in the world which performed at such speed! This is how Kartsev inspired and captivated the entire team. The M-9 complex draft was already designed in 1967. At the when legending it of the Ministry, it was given a positive estimation.

Unfortunately, the M-9 complex was not mass-manufactured, though its design and the successful use of the model demonstrated the powerful creative potential of Kartsev's team. In 1967, a special year for the M-9's designers, the Scientific Research Institute of Computing Complexes (SRICC) was founded. Kartsev's department became its operational center and Kartsev himself was appointed a director. It was official acknowledgment of Kartsev's scientific school.

In 1969, the government issued an Order on creation of the M-10 electronic computer, based on M-9 the vector numeric computer.

Accoding to L.Ivanov, doctor of Technical Sciences "The event was preceded by an authoritative meeting on prospects of two already started projects: the "Elbrus", managed by academician Lebedev and the M-10 managed by Kartsev. Lebedev argued vigorously against making "Elbrus" as a multi-processor computer insisting on a single processor version for achievement of maximum productivity. Academician Glushkov supported both versions, and they both were approved". (see "Radio-electronics Problems" magazine, Iss., 1993). The same year, the development of the M-10's engineering documentation began and soon the specifications were sent to the manufacturer. In 1970, the manufacturer began preparing an experimental model for production of. By the middle of 1970 all construction documentation had been presented to the manufacturer and within a year, the manufacturer had already prepared the first experimental M-10 model for adjustment. At the same time, industrial computer models were undergoing corrections of its engineering documentation and preparation for manufacture. That was a very hard year for Kartsev. The intensive work took its toll on him. An extensive heart attack put him in bed for several months. Fortunately, everything turned out well.

Before July 1973 all units of the first model had been assembled and tested according to technical specifications. They were then assembled for a complex adjustment of the whole computer. In September of the same year the first industrial model of the M-10 successfully passed all tests and the experimental operation for software debugging started.

By December, the second computer testing had been finished and its mass-manufacturing began. Its industrial production went on for more than 15 years. Dozens of computers were produced, most of which have been currently still under operation. On the basis of M-10 computers, a number of powerful computer complexes were developed. In 1976, such M-10 computer complex successfully passed state tests (both hard-and-software).

In order to explain the importance of the M-10's creation, something about its designated purpose should be noted. The M-10 was a secret for a long time because it was developed for the Missile Attack Warning System (SPRN) and for general outer space scanning. The system was designed to give the country's leaders comprehensive information on feasible threat of missile attack as well as to provide constant observation of outer space. So, the system had mostly defensive purpose. At present, there exist in near-Earth orbital space about 17 thousand objects of various origin, including functioning and dead satellites, parts of missiles, etc. The first level of SPRN was a space one: satellites detect the launch of missiles by their jets. The frame of the system - its second on-ground level consisted of powerful radar-tracking stations located on the border of the country (before the collapse of the USSR, there were nine: near Riga, Murmansk, Pechera, Irkutsk, Lake Balkash, Mingechaur, Sevastopol, Mukachevo). In addition, there was a network of M-10-based computing complexes.

In our time, even ordinary training of troops is connected with the launches of various classes of missiles. And what would happen if nuclear missiles got into the hands of political adventurers, arrogant "chiefs" capable of organizing so-called "unsanctioned" launches? Fast and exact evaluation of the situation is reguired; otherwise, the consequences could be suicidal for the whole planet. That's why there appeared the paradoxical at first glance idea of world-scale coordination of rocket attack warning and outer-space watching systems.

It should be noted that requirements for such computer systems are extremely high. The path of a launched rocket is measured in seconds. At the same time, the information flow from radar-tracking stations is huge.

Until 1980, the M-10 had the USSR's highest performance values (according to some estimates, it equaled 20-30 millions op/s), storage volume and data transmission speed of multiplex channel. For the first time in the world, a whole series of progressive new tasks were solved including the computer network provided for direct information interchange and synchronous coordination of up to 7 computers (bypassing a multiplex channel) and dynamic sharing of hardware between separately programmed computers; realization of an automatic reconfiguration of a field of processors; implementation of a second level internal 4MB RAM storage with random access; external exchange between both levels of internal memory.

The Last Battle

In 1978, Kartsev, using the experience acquired in the development, manufacturing and operation of the M-10 and M-10M computers and modern achievements in technique and technology, proposed development of a multi-processor vector computer. They decided to call the new computer the M-13.

In 1979, the development of engineering documentation started. The plants-manufacturer were selected. During 1980-1981, the engineering documentation for the separate units was delivered to the manufacturers.

The M-13 became the fourth generation computer. Used Large integrated circuits were used as an elemental base. The main purpose of the computer was real time processing of large information flows. Its architecture used four basic units: a central processor, hardware equipment for operational system support, remote control interfaces and a special processor.

The M-13's special processor was used for processing large arrays of relatively short digit-length information (fast Furie transformations, correlation function calculations, comparisons with threshold, hypothesis testing etc.) The processor used the production of two complex numbers (two points Furie transformation) as base operation.

The equivalent operational speed of the M-13's special processor with maximum hardware performance reached 2.4 billion op/s.

One of Very Few

All his life Mikhail Kartsev demonstrated great creative activity. His books on the theory of arithmetic units and computer structure design became basic textbooks for computer designers. The special computers designed under Kartsev's management and used by Soviet Army troops were less known. The M-4M computers (army codes 5E71, 5E72, 5E73) were ten times more powerful than their contemporary models M-220, BESM-4 and others. They operated at military facilities from 1967 to 1981, and were mass-manufactured. Their failure-free operating time was about 700-1000 hours.

The M-10 computer (army code 5E66) significantly exceeded contemporary domestic models (BESM-6, EC-1060).

The USSR's largest multi-computer complex was built using 5E71-73 computers. In it, 76 computers functioned on a common algorithm and were used in round-the-clock operation, connected by data transmissions channels over distances of tens of thousands kilometers.

For the first time in multi-processor fourth-generation systems there came into use: equipment for operational step-cycles (allowing for program independence from a number of system processors), equipment of segment-page storage organization (which enhanced the possibilities of file systems), and software control processors for handling operations of Furie, Adamare, Walsh, and Fresnel transformation types, calculation of correlation functions, space filtration problems, etc. The average operating speed of the central processing unit was as high as 50 million op/s (or up to 200 million short op/s), with internal storage of 34 MBytes, external information exchange transmission speeds up to 100 MBytes per second, and an equivalent operational speed of peripheral processors of up to 2 billion op/s.

M.Kartsev is an author of fundamental theoretical works on computer technology (5 books, 55 articles and reports, 16 invention certificates). His books "Arithmetic Units of Electronic Digital Computers" (Russian edition 1958 and later published abroad) and "Digital Computer Arithmetic" (1969) provided the theoretical base for arithmetic units and its conclusions have been widely used in textbooks. His last works, "Digital Computer Architecture" and "Computing Systems and Synchronous Arithmetic" (1978) were actually the first attempt to establish a scientific base for computer structure and parallel calculation equipment design.

Kartsev was awarded the Lenin Order (1978), the Red Banner of Labor (1971), the Order "Medal of Honor" (1966), a medal "For Valiant Labor" and the State Prize of the USSR (1967) for his work and achievements.

In 1993, the Institute he founded was renamed as Kartsev Institute of Computer Complexes.


Medal "Computer Pioneer" IEEE Computer Society Laureates