From first-generation computers to supercomputers
I started working at JINR in 1963 but for the first time, I had been at the Institute a year earlier - this is how the circumstances developed. In my life, in general, many important turns happened seemingly by accident but then it turned out that the accident was very useful. I graduated from school in 1957 with a gold medal that in those times allowed me to enter any institute without entrance exams. I didn't know exactly which institute to choose: I was thinking about Moscow State University, MSTU named after N. E. Bauman, MAI. My school friend offered to enter MEPhI and made a weighty argument, "They have a sports camp on the Black Sea". And at school, I was seriously engaged in gymnastics, had a rank, performed at competitions. The argument turned out to be weighty for me. At MEPhI, from three faculties - theoretical physics, energy and computer engineering I chose the latter. There was a feeling that this area was just beginning and the future was behind it.
In senior courses, higher mathematics was taught by Evgeny Petrovich Zhidkov. At JINR, he was Head of the Department of Computational Mathematics and Counting Machines of LTP. I lectured easily, with a smile and I was usually one of the first to go to the exam for it. Evgeny Petrovich knew my abilities. He pushed the ticket aside and like a magician, took out a card with tricky questions from his chest pocket. As a rule, I managed to answer them correctly and I got "excellent". When the time came for pre‑graduation practice and diploma, it was thanks to E.P. Zhidkov that I was invited to JINR.
I immediately liked Dubna: a lot of greenery, some kind of special openness, interesting people around. In the hostel, one of my roommates was the graduate student Vladimir Georgievich Kadyshevsky that was preparing to defend his thesis. Since then, we have established very warm, friendly relations with him. In the hostel, I met Igor Nikolaevich Silin. Usually, he had a large company in his room to listen to recordings of songs by Bulat Okudzhava. The living atmosphere, interesting conversations and disputes, songs, youth were kept in my memory. After having defended with a "red diploma" in June 1963, I was hired by the JINR Computing Centre reorganized from the LTP department into an independent unit. My work began with the operation of the Kyiv computer. The machine had more than four thousand lamps and was quite capricious. The repair service, where I was sent to, kept busy.
The RAM of the Kyiv computer was about 4 KB, the external memory on a magnetic drum was less than 16 KB and the performance was 10‑12 thousand operations per second. Today, these characteristics cause a smile but then, in such memory they met real problems. You can imagine how much ingenuity was required from programmers and mathematicians, especially since they wrote then in machine codes. There was also a M‑20 computer, also on lamps, about twice as powerful and significantly more reliable. How much perseverance and patience it was necessary to have in order to debug programmes and to get important results on these computers.
In 1966, JINR Directorate made a strategic decision to found the Laboratory of Computing Techniques and Automation (LCTA). Mikhail Grigorievich Meshcheryakov was appointed Director of the new Laboratory, Nikolay Nikolaevich Govorun and Georgy Ivanovich Zabiyakin were his deputies. Mikhail Grigorievich immediately set two big tasks: to equip the laboratory with modern computing equipment and to build a separate building for LCTA, where it would be possible to assemble a staff, to place computing power and an automation equipment. Looking back now, it's obvious how deeply thought-out a strategy it was. To meet the first problem, in particular, a contract was signed with the American Company Control Data Corporation for the purchase of CDC‑1604A computers. At that time, it looked like an almost insoluble problem: the United States categorically prohibited the sale of computing equipment in the USSR and other socialist countries. But with the efforts of JINR and LCTA Directorate, obstacles were overcome. The machine was very reliable; its mini variant was used even in the American submarine fleet. Under the terms of the contract, two engineers and one programmer were to be sent to the Company's courses in Germany to study the machine and software for its further operation at LCTA.
I knew English and computer terminology well, so my candidacy also arose. I was just going to go into the mathematical department to E. P. Zhidkov, I wrote a statement. Mikhail Grigorievich signed it, putting a sweeping positive resolution that went from the bottom corner to the opposite top. It was needed to take the application to the personnel department. And at that moment, N. N. Govorun intervened in my fate. He said something like this, "Evgeny Petrovich can wait. There is a contract and the opportunity to go to Germany. To listen to courses, to talk, to broaden your horizons and life experience... As you know, it was impossible to resist the convictions of N. N. Govorun and I agreed. But my consent alone was not enough. Going to a capitalist country was serious business then. The candidate for the trip had to be a member of the CPSU, family, to have experience of traveling to social countries - in a word, to confirm that he is a reliable Soviet citizen. I had none of the above. But if Nikolay Nikolaevich decided something, he necessarily sought. I think the indisputable authority of N. N. Bogolyubov in higher instances also helped. As a result, my candidacy was approved with a feature in which the necessary stamp was "politically competent, morally stable, modest in everyday life". The second engineer was sent by V. P. Mirolyubov, as a programmer - G. A. Ososkov.
In 1968, CDC-1604A was obtained by LCTA. It was the first American computer legally installed in the USSR. I joined the team of engineers for its operation to transfer the experience gained in courses in Germany. Its performance was about 100 thousand operations per second with 192 KB of RAM. The computer had a translator from Fortran and other programming languages. In the same year, a BESM‑6 computer was installed in LCTA with a capacity of about 1 million operations per second. Later, the team of N. N. Govorun developed a translator from the Fortran language for this computer. Today, it may seem strange and unusual but in those years, there was the concept of "engineering machine time". Every morning, in order to pass the computer to users, the engineers set it up for two or three hours: they ran tests, identified unreliable components and fixed problems. The CDC‑1604A differed in being completely semiconductor and so reliable that it could operate around the clock without morning debugging. Therefore, CDC‑1604A engineers took away a couple of morning hours for development, including the organization of communication lines with other computers. For example, I participated in connecting a graphic display with a light pencil to this computer, developed at LCTA and in developing software for it. Today, displays and interactive work with a computer seem to be something taken for granted. Then, in the USSR displays for computers did not yet exist.
The acquisition of the CDC-1604A computer opened the way for new contracts. A significantly more powerful CDC‑6200 computer was delivered that was then modified to older models of the CDC‑6xxx series. It was also very important that within the framework of these contracts, modern tape readers were purchased, not only for LCTA, but also for other laboratories. Prior to that, if the information was registered on a Russian tape recorder, it was guaranteed to be read almost only on it. Transferring data from one computer to another was an almost insoluble task. With the CDC technique, magnetic tapes could be rearranged from a tape recorder to a tape recorder, to be carried and successfully read on any compatible hardware. The available CDC tape recorders on computers of other JINR laboratories allowed to meet the problem of information exchange between laboratories and the central computer complex LCTA.
A new area in the development of computing at JINR was the occurrence of middle‑class control computers, such as the small M‑6000 computer that was then widely used in Soviet research institutes and organizations. In LCTA, a graphic display station was M-6000 developed on the basis of a computer. My group developed handy graphics software for it. Thanks to it, the station was later used in various problems of theoretical and experimental physics. The station's basic graphics software was implemented in dozens of research organizations in the country.
In 1965, IBM began production of the IBM/360 series. All machines in this series had a unified interface for external devices and were software compatible with each other. I have done a lot of translations of technical documentation and popular publications on this series. Several books have been published with my translation. I was even invited to the post of head of department in the newly established Research Centre for Electronic Computing Technology (RCECT) in Moscow. It was organized to coordinate the work of the socialist countries to develop a Russian series of EU computers - an analogue of IBM/360. But I chose to stay in Dubna. There were interesting tasks here, a comfortable environment, which I have already attached myself to.
The ending follows.
Olga TARANTINA,
Photos from the archive of Alexander KARLOV