Mikhail Alexanderovich Kartsev

Home → Hall of Fame → Mikhail Alexanderovich Kartsev

Composed and translator
by Alexander Nitussov

M.A. Kartsev was born in Kiev into a pedagogic family, on the 10 th of May 1923. Mikhail belonged to the generation of young people who merrily celebrated their graduation from secondary school in the evening of the 21 st of June 1941 (a day before outbreak of the Nazi invasion) and the next day were already preparing themselves for the years of hardships, losses and bloodthirsty war battles. He was mobilized to army in the October 41 st and served until the February 47. Being enlisted to a tank regiment he fought with it in the North Caucasus, Ukraine, Rumania, Hungary, Czechoslovakia and Austria.

Kartsev was lucky to pass through the battles without suffering serious harm, and soon after demobilization he entered the Radio faculty of the Moscow Power Engineering Institute (MEPI) [1]. At the third university year he passed all exams also for the fourth year [2], and in the 1950-1951, being a student of the fifth year, also worked at the laboratory of electric systems of the Power Engineering (research) Institute at the USSR Academy of Sciences. There he was engaged into a new project - design of “digital electronic computing machine M-1”, one of the USSR first two computers. The team of young engineers – all graduates of MEPI - was headed by academician Isaak Bruk and Nikolay Matykhin was its “executive designer in chief”. Matyukhin was four years younger than Kartsev himself.

In 1952 Kartsev received a permanent position at the Bruk's laboratory as a head of new project – electronic computer M-2. The small group managed with it within a year and half, and in 1953 M-2 was already in operation. Its main characteristics were comparable with the contemporary analogues “Strela” and BESM 13. Unlikely to most of computers of that time (all over the world) Brook's series “M” were purely civil scientific-aimed machines. For several years those three were the basic machines at of the soviet computer centres. M-2 has been in operation 15 years long. It was almost never turned off.

In the winter 1954-1955 Kartsev basically modified M-2 and, in 1956, he introduced a ferrite memory device with total capacity 4096 words of thirty four decimal places. That needed additional register for memory sectors switching during a program performance. While modernising the computer Kartsev also realised the idea of truncated addresses and operation codes, what turned into a prototype of the executive addresses forming principle implemented in computers of the second and third generations.

In 1956 Kartsev submitted his dissertation thesis on basic principles of arithmetic devices in digital computers, that was awarded the “candidate of sciences” degree. Results of his research were published in Moscow under the title “Arithmetic devices of electronic digital computers” in 1958; some years later the book was translated and published abroad.

In 1957 I . S . Bruk started one more project - M-4. It was an electronic computer for control system of a new experimental radiolocation complex, that was being designed at the Institute of Radio-Engineering AS USSR by academician A.L. Mints. Bruk appointed Kartsev a scientific head of the whole work . Since the first soviet transistors had already been put into commercial production, Kartsev decided to use them as basic components for the computer circuitry. Thus M-4 appeared among the first computers of the second generation. In April 1958 the USSR Council of Ministers nominated the plant for M-4 serial production. That was electromechanical plant in Zagorsk (near Moscow ). A.G. Shishlov was the director deputy on engineering part and V.S. Semenikhin was the chief of special design bureau in charge of the project. Later Semenikhin was elected an academician and headed design of computer complexes for the state defense programs.

Working on M-4 Kartsev became one of the first to separate data memory of RAM from the memory for programs and constants; the later were stored in ROM unit. Later in the USA, analogous principle was named “Harvard Architecture”. Those structural modifications were aimed at increasing of performance reliability, by lowering general danger of failures in the (defense) systems based on M-4 computers. In April 1958 Bruk's institute consigned the plant complete set of documentary for M-4. The first two computers were ready in 1959. By the end of 1960 one of them was fully tested and all necessary adjustments completed. It was handed over to the Institute of Radio Engineering . In July 1962 the complex testing of M-4 coupled with the device for preliminary information processing and connecting to object at the experimental radiolocation system was successfully completed. M-4 was put into serial production by the government directive of November 1962.

Scientific school of Kartsev accumulated and implemented the M-4 experience for subsequent simplification of the “bringing into production” procedure with new computers. That permitted more efficient development without ordering big operating models of worked-out machines at the manufacturing plant. Such approach saved them reasonable amount of time and money but needed correct and precise technical solutions and, what was even more important, daring and new design ideology. That was also a challenge for the -still young- scientists since they had to consider all realities (and weak points) of manufacturing plants on the project drafts and other documents preparation stage.

In spite of the issued directive on M-4 serial production, Kartsev insisted on its immediate improvement considering the progress in electronic components. Modernisation increased its speed almost ten times . In 1964 new computer M-4M was completed. Its speed reached 220 thousand op/sec. M-4M was very famous machine, or rather a multiprocessor complex, and its production lasted until 1985. More then 100 of them were mounted for defense purposes.

Years later Kartsev commented it in the following words, ”25 years ago, in 1957, we began work on one of the first transistor computers M-4. It operated in real time mode and satisfied all testing demands. Although it was recommended to serial production by a government decision we could see that it was far from being completed. That was the first transistor computer. Its adjustment and regulations were difficult and it could be hardly repeated by manufacturers. Besides that, the period of 1957-1962 brought much progress into semiconductor development, what enabled improvement of the computer almost ten times. It could be also ten times more powerful then other Soviet computers. Fierce discussions lasted the whole winter 1962-1963. The institute administration opposed our proposals with all possible arguments, insisting on utopian nature of the new project. Luckily the decision of the Industrial and military commission of the USSR Council of Ministers put end to that arguing and in the same month we handed over the first set of project documents to the plant. It was the central arithmetic unit. By the August 1963 the plant received the rest of it and in August 1964 the first two new machines were ready for testing. In October 1964, less than two years after the governmental directive, two complexes were already on destination cites and in operation and in December 1964 five more were shipped to the customers”.

Computers M-4M and M-42M, enlarged with peripheral processing device and modems to couple them with communication channels, formed basis for multi-computer computation complexes joined into a powerful computation system operating in a real-time mode.

Kartsev's fundamental researches on M-4M brought him a doctorate, and some of his collaborators received the degree “candidate of sciences”. In 1967 he was awarded the USSR State Premium (one of the highest awards).

In 1967 Kartsev presented a new daring project – computation complex M-9 with performance of about 1 milliard (billion) op/sec. It was the time when BESM-6 - one of the best computers of that period, with 1 million op/sec, was almost completed. His ideas of the computer complex architecture were based on absolutely new principles. Kartsev suggested that some types of computation problems needed computers that performed operations not upon numbers but functions of one or two variables, that should be represented with discrete sets of values. Such set of operations upon functional space should be realized by means of a matrix of 32x32 arithmetic devices, what he named, “functional cluster”. Arithmetic devices should operate from the same clock-rate generator performing every operation within one or two cycles and providing data exchange between every arithmetic unit and memory unit of the cluster both at the end of each operation and in the beginning of the next one. At present it is the basic principle of the RISC architecture. A line of 32 arithmetic devices, “number cluster”, was intended in M-9 for operations with multi-dimensional vectors. Operations with data arrays should be performed by the “associative cluster”. That was his idea of the high-level parallel computations operation with various data, for broad spectrum of computation problems.

Unfortunately those progressive ideas of his, much ahead of their time, were not realized on practice that time, first of all for the lack of necessary electronic components.

Computer M-10 was the next project, started in 1969. It was a vector machine with a broad spectrum of microelectronic elements and the dominating idea of “numeral clusters” transferred from the theoretical project of M-9. It had been essentially prepared by the previous project, so that time the work progressed quickly, and already in 1973 the first industrially manufactured set of that large multiprocessor complex was successfully operating. M-10 was in serial production for about 15 years. Several dozens of those complexes have been produced and widely used for the defence purposes, especially in the systems of missile assault early warning. M-10 was a computation system of synchronous type and belonged to the computers of third generation. It was also the basic machine for the neighbouring space monitoring network formed of numerous radars and communication channels.

M-10 was the first computer developed in new building of the –new- Institute of Computing Complexes . For more than ten previous years Kartsev's team (still as a department of the –Bruk's- Institute of Electronic Control Machines) had no special single place to work in, so they had to be spread over several sites in various parts of Moscow (some in semi-cellars). Only in 1967 special buildings were constructed for them .

Serial (commercial) production of M-10 could be also estimated as double triumph. Not only the ideas of M-9 were finally proved on practice. In the beginning most of the contemporary experts were more than unbelieving, what, besides negative emotions, had negative influence on financial side of the project. “Many sceptics insisted that it was an absurd idea… absolutely unrealistic, and we ourselves were insane…”, remembered Katrsev afterwards. As for the attitude to complex structures combining two, three or more M-10s, it was negative in the same proportion. However, having appeared in serial production the computer brought the authors 18 patents and also the USSR State Premium, in 1977.

M. Kartsev was fully aware of the computer special importance also for the scientific researches in general. There's no need to say that the powerful defence equipment was designed and produced in conditions of high secrecy, so he had to make energetic efforts to promote the idea of open publications on M-10 (or at least on “half-secret” [4] status to be accessible for “civil scientists”). Finally that was also a success, thus he initiated very complex calculations on the physics of solid matter (several dozens of times quicker than on contemporary civil ES-1040), also on plasma modelling, what was 20 times quicker than with BESM-6 (one of the world's best), provided that the whole problem was stored in BESM-6 RAM). It was also M-10 that could successfully model and bring authentic data on the plasma collapse. The parallel attempts made in the USA with computer CDC-7600 have been a failure.

In 1978 Kartsev began his new project – multiprocessor vector computer M-13 based on big integrated circuits. It was the USSR first vector-conveyer computer. Central part of M-13 consisted of “ensemble” (Kartsev's definition) of 225 processors, which automatically “distributed” performance of arithmetic and other operations, subroutines, etc. among themselves all time during the computer operation. Similarly to M-10, M-13 was also “in charge” of the whole neighbouring space monitoring.

Besides the computer projects Kartsev also was engaged in theoretical researches. He authored several books of principle importance on computation arithmetic and foundations of computer architecture. Similarly (and almost simultaneously) with researches of S. Lebedev he published book “Computation systems and synchronous arithmetic” ( Moscow, 1978) which was presenting scientific principles of structural design for parallel computation systems.

Production of M-13 was also successful, beginning in 1984, unfortunately he would not see it. M.A. Katrsev died 23.04.1983 in Moscow .

For his numerous successful projects Kartsev was awarded the order of “Red Banner of Labour”, the “Lenin order”, order the “Honourable Sigh” for merits, order the “Red Star” [5] and numerous medals.

Ten years after his death the institute Kartsev founded was named after him.

Notes

1.One of the biggest technical universities in Russia /USSR/

2. What was an exception, for all Russian universities (unlikely to many modern Western ones) had (and chiefly have) strictly prescribed curriculum schedule and very intensive educational plan. The big volume of obligatory work left little time for “studying in advance”, what therefore needed really unusual abilities.

3. The both were produced by much more powerful scientific enterprises. STRELA was the first serially manufactured computer, ordered mainly by the ministry of defense and BESM was created by academician S. Lebedev – one of the world leading computer scientists and the director of an academic computer research institute.

4.Several grades of secrecy ranged from the strict “top secret” to rather “liberal” “for the service purposes”, or “for the concerned personnel (civil) only”.

5.Not to be confused with the “Gold Star” the USSR highest order.