The birth of the Soviet missile defense system. BESM. Saga
Under Stalin, Lebedev's career did not work out, in the process of designing MESM, as we already mentioned, he almost got a hat for daring to promote smart Jews to work, turning a blind eye to their race, BESM did not go into the series, but under Khrushchev Lebedev turned around with might and main. In 1953, he became the director of ITMiVT, in 1956 he received a Hero of Socialist Labor, despite the secrecy of his work, Lebedev went out to the USA (where, by the way, he talks more language than in the USSR, as a result, Americans publish in magazines stories about the secret computer SKB-245 "Volga", which almost no one has heard of even in modern Russia), does not officially participate in any political battles, nevertheless, his beloved student Burtsev receives a contract first for the construction of the M-40 / M- 50 for the "A" system, and then "Elbrus" for the A-135. What is especially incredible: the M-40 was built by Burtsev, the A-35 was used by Yuditsky's K-340A, and the Lenin Prize for the development of missile defense in 1966 was received by ... Lebedev! There is no need to talk about the fact that he was a full academician of both the Academy of Sciences of the Ukrainian SSR and the USSR.
In general, the awards poured on him more than on Shokin - the Stalin Prize, the Lenin Prize, the USSR State Prize, the Star of the Hero, four Orders of Lenin, two Orders of the Red Banner of Labor, and on top of the cherry on top was the Order of the October Revolution. Posthumously, he also received the highest honor in the field of computing - the Computer Pioneer Award, aka the Babbage Medal from the IEEE Computer Society discovered the Minsk computer and awarded its creators G.P. Lopato and G.K. Stolyarov).
At the same time, Lebedev, unlike many, was not noticed in very straightforward and dirty intrigues, he did not sit around and put anyone in the open, and did not get dirty in those high-profile scandals around missile defense that we described. However, since 1953, he managed to systematically win wherever he took part, carefully squeezing out all competitors of ITMiVT from everywhere. Even for the A-135 system, at first they wanted to use Kartsev's machines (and they would have been quite capable!), But Lebedev was a fanatic of single-processor systems (and a patriot of himself and his institute) and at the meeting offered the job of his student Burtsev - Elbrus-1. Moreover, he proposed, as usual, not aggressively, delicately, but at the same time even Academician Glushkov did not dare to insist on his own and peacefully agreed to build, in addition to the M-10, Elbrus. And so it was with absolutely everything he undertook. Posthumous glory was also given to him in full measure, right down to praise in the spirit
Lebedev's developments determined the main road of world computer engineering for several decades to come,
as the President of the Russian Academy of Sciences (1991–2013) Yu.S. Osipov.
Well, let's see what they determined and in which direction. We have already written about MESM and BESM separately, so we will not dwell on the details of their design, we will talk about comparing them with Western architectures and about stories ITMiVT.
Once Sergei Alekseevich sketched in our magazine an electrical circuit based on a triode with a transformer with 6 windings, three semiconductor diodes, a storage capacity, and two resistances.
This is how one of the designers of MESM, Igor Mikhailovich Lisovsky, recalled his work with the chief of the domestic computer industry. This scheme (quite successful from the point of view of the 1st generation of machines), in general, was Lebedev's main electrotechnical achievement, and it was this scheme that he repeated in one form or another in the entire BESM series, without switching from transistors and diodes on microcircuits until death.
By itself, the MESM was, in principle, not bad, but since such an application is direct world and straight pole, then we will ask from it in full how much its design would be competitive in the world market - let's compare the MESM, well, at least with UNIVAC I released at the same time.
Not a single article is complete without solemn mentions that MESM was the fastest, best and first in continental Europe, and this is so, that's just ...
She had no one to compete with there at all. Eastern Europe lay in monstrous ruins, Germany was occupied by the USSR and the allies and divided into zones, France before the war generally dealt only with tabulators, in Spain and Italy there was not even that.
In general, with whom MESM was to compete is decidedly incomprehensible, except perhaps with the relay Z4 of Konrad Zuse from ETH Zurich. Naturally, the relay machine, in principle, could not overtake the electronic one, the speed of the Z4 was 20 times less (and at the same time the power consumption and size), nevertheless it was quite enough to, for example, design the Swiss jet fighter P-16.
As for the world, by the time the MESM was put into operation, Manchester Mark 1, EDSAC, BINAC, CSIRAC, SEAC, SWAC, Univac 1101, ERA Atlas, Pilot ACE, Harvard Mark III, Ferranti Mark 1, EDVAC, Harwell Dekatron Computer, Whirlwind, LEO I, Hollerith Electronic Computer and IAS machine.
An outraged reader can refer to the fact that these computers were made by the bourgeoisie, the British and the Americans, and not the unfortunate sufferer of the USSR, but, sorry, England got no less.
During the two battles for Britain, despite the advanced air defense (without it, England would have been turned into a post-apocalyptic wasteland), a huge number of factories, factories, railway junctions were destroyed, docks and port facilities were destroyed. In cities, including the capital, entire areas were wiped off the face of the earth. Housing losses amounted to more than four million homes. Every third house in the UK was either destroyed or rendered uninhabitable, with the result that a quarter of the country needed a roof over their heads and huddled in subways and basements. The people were starving, it got to the point that the authorities were distributing cartridges for free to those wishing to shoot the squirrels that had bred during the war years. The cards for bread were canceled only in 1948, for gasoline - in 1950, for sugar - in 1953, and for meat - in 1954. So from this point of view, we have full parity.
Children miners dig coal, 1943. Firefighters unsuccessfully try to extinguish the burning ruins of London after the German raid aviation... Ordinary life in England in the 1960s – 1070s is no better than the barracks and communal apartments of the USSR.
Photo flashbak.com
and New York Times Paris Bureau Collection
Photo flashbak.com
and New York Times Paris Bureau Collection
In 1953, a large version of Lebedev's architecture appeared - BESM. At this time, Remington Rand 409, Harvard Mark IV, Max Planck Institute G1, ORDVAC, ILLIAC I, MANIAC I, IBM 701 and Bull Gamma 3 were operating in the world, so it is also difficult to consider it as something unique. But let's make a comparison.
From a technical point of view, the BESM is a rather primitive machine, compared to the 701, it is more bulky and consumes more energy. From a performance point of view, it also does not make a devastating impression. Well, from the point of view of influence on the world computer industry - even more so. As you can see, MESM and BESM were quite adequate by the standards of the USSR, but they were not drawn to the role of a world flagship.
Let's see how the ITMiVT was founded and the following Lebedev machines were built.
14 days after the atomic bombing of Hiroshima, by the resolution of the State Defense Committee No. 9887ss / op dated August 20, 1945, signed by I.V. Stalin under the State Defense Committee, a Special Committee was formed to manage all work on the use of atomic energy. Soviet computerization grew out of the Soviet nuclear project - from the very beginning of the development of the bomb it became obvious to all participants that they would have to count and count a lot.
At the same time, in 1945, the situation in the USSR was, to put it mildly, rather tight with counting equipment.
Since 1931, the USA produced the IBM 601 Multiplying Punch, a multiplying tabulator, one of the most advanced calculators of those years. He read two multipliers up to eight characters from a punched card and knocked out their product in an empty field of the same card, in addition, he knew how to subtract and add. In addition, it allowed for a significant expansion of the functionality and rather complex sequences of calculations.
For example, the tabulator that was delivered to Wallace John Eckert's Astronomical Laboratory in 1933 was a special model developed for Eckert by one of IBM's lead engineers at Endicott that was capable of interpolation. Eckert went even further, in 1936, pairing it with an IBM 285 tabulator and an IBM 016 duplicate puncher using a computation control switch of his own design, producing the first machine that automatically performs complex scientific computations.
By 1946, the range of IBM tabulators was extremely extensive, such as the Type 602 Calculating Punch, the most sophisticated and sophisticated, adding, subtracting, multiplying, and dividing at 100 cards per minute and controlled from the software panel. Books of the 1940-1950s. full of panel connection diagrams of this machine, which allowed for complex calculations, for example, Mass Spectrometer Calculations on the IBM 602-A Calculating Punch, WH King Jr., William Priestley Jr.
In the same 1946, the IBM Type 603 Electronic Multiplier appeared - the world's first electronic calculator with 300 lamps, with a speed of 0,017 seconds. by multiplication. His successor 604 had 1 lamps and could do all 100 actions. In 4, IBM connected it to a 1949 Electronic Accounting Machine and an additional relay memory unit in the spirit of Eckert's work and received a Card Programmed Calculator (CPC), more than 402 units sold in total (an advanced version from a bundle of 2 and 500 (or 605/412 ) + Type 407 Electronic Storage Unit released in 418 as Model A942, aka CPC-II).
We do not mention here full-fledged relay and lamp-relay computers, both developed by IBM (PSRC, SSEC, ASCC), and Bell Labs (Model I - Model V), and Harvard (Harvard Mk II, Mk III / ADEC), but also, of course, ENIAC and Vannevar Bush's analog differential analyzers (they carried most of the computational load of the Manhattan project).
In addition, in the 1940s, the United States was the first country in the world to provide electromechanical adding machines, an extremely advanced type. They were produced by 4 famous companies - Burroughs, Friden, Felt & Tarrant and Victor Adding Machine Co. (plus a limited number of no less advanced Mercedes cars were imported from Europe).
What of all this splendor was there in the USSR at the beginning of the atomic project (well, except for slide rule)?
There was Brook's integrator, a simpler model than Vannevar Bush's machines. In the area of tabulators and adding machines, everything was not just gloomy, but very gloomy. In August 1923, the USSR State Planning Commission was created under the USSR Labor and Defense Council - the legendary State Planning Committee. The turn of the entire economy towards total production planning - from nails and toilet paper to tanks, immediately gave rise to the need for total computerization, even surpassing the American one. It was unrealistic to count everything by hand, statistics bodies and planning commissions filled up higher organizations with applications for calculating machines. In June 1926, an employee of the Supreme Council of the National Economy of the USSR (VSNKh), Nightingale wrote:
The production of adding machines was delivered at the Suschevsky plant ... This plant produces up to 100 pcs. per month. This quantity satisfies about 30% of the existing demand, which can be determined at 400-500 pcs. per month. In the first half of the year 25/26, 123 rubles of calculating machines were imported from abroad. According to the reviews of the Main. There. Offices, adding machines are one of the most cost-effective commodities that come from abroad through postal parcels and smuggling.
Pay attention to the word "contraband"!
A striking situation is a total contradiction between the requirements of the system (to catch up and overtake, implement and improve) and the inability of the system to provide a resource for fulfilling its own requirements. The executors themselves: accountants, bookkeepers, researchers were forced to secretly obtain equipment for government work at their own expense by illegal means.
The real need of the country was about 10 thousand adding machines a year, while at least 80-90% of calculating machines passed by customs. Finally, the obvious delirium of the situation was realized, and a resolution was adopted to establish the production of our own adding machines, declaratively setting humane prices on them for the victory of smuggling in an economic way, but this was not easy. Glavkontsesskom concluded:
Nowadays, the adding machines of the Sushchevsky plant are in many ways inferior to foreign ones: they are outdated in their design, cumbersome, unproductive and greatly tire those working for them. Prices for adding machines of our production are 65–175% higher than retail prices ... The reasons for such a high level of prices ... must be sought in the obsolete and worn-out equipment and the discrepancy between organizational and production processes at the plant.
Odner's adding machines (in the famous version of "Iron Felix") were nevertheless produced in the USSR until 1978 (!), Although they were monstrously primitive and inconvenient. If it was impossible to correct the situation in fact, socialism could always teach how to correctly and partisanly comprehend it, for example, in 1949 at the exhibition "Socialist Accounting" in Moscow, they wrote about the same adding machine:
The stand "Domestic Inventors and Designers of Calculating Machines" in the opening hall of the Exhibition introduces the works that were of exceptional importance for the development of calculating machine building, which established a national priority in the creation of many basic designs of calculating machines and instruments. The introductory part of this section demonstrates an original copy of the adding machine, produced in the 80s of the last century by the St. Petersburg mechanic Odner, who created the first design of a computer that was widely used. Odner's adding machine is a prototype of all existing adding machines, both domestic and foreign.
Well, there is something to be proud of, although what relation does the scheme of the Swede Odner (Willgodt Theophil Odhner) to the engineer of the Swedish company Ludvig Emmanuel Nobel, produced by the English-Swedish factory Odhner & Hill Frank N. Hill, have to do with great Russian inventions, it is absolutely incomprehensible, just as it is unclear at what time it suddenly became the basis for the development of many times more advanced push-button electromechanical machines Burroughs, Friden, Felt & Tarrant and the like that are completely different in circuitry, in comparison with which the poor Odner looked like "Zaporozhets" against the background of McLaren.
In 1927, to the tenth anniversary of the revolution, the Moscow plant. Dzerzhinsky began production of the most famous Odner's clone - the Felix adding machine, the same one that was produced until the end of the 1970s. In 1935, the USSR produced a keyboard semi-automatic adding machine KSM-1, an analogue of Comptometr, but simpler. In the post-war years, KSM-2 semiautomatic devices were produced (with slight differences in design from KSM-1, but with a more convenient arrangement of working parts).
Burroughs vs. Iron Felix. Choose the most ergonomic option. Photo vintagecalculators.com and habr.com
Nevertheless, the keyboard machines turned out to be too complicated for the USSR, and we did not have full-fledged analogs of Burroughs. The maximum development of this technique was the unique Vilnius calculators, made on the basis of electric meters, relays and semiconductor diodes. Indeed, this was the missing link between electromechanical and electronic machines, and no other model of a compact relay computer is known today, because in the West they immediately switched from mechanics to electronics.
The rarity was developed by the Kirov NIISVT in the first half of the 1960s, and the production was launched approximately in 1965 in the Volga-Vyatka Economic Council under the Vyatka trademark and in 1967 at the Vilnius Electric Meter Plant under the Vilnius trademark. There was an advanced modification of 1968 with an optimized design without a bunch of parts, but they were produced for a short time, only until 1969.
In addition to Odner, more complex designs were cloned in the USSR (of course, without a license), for example, the Facit TK push-button machine turned into VK-1. The manufacturing quality of these more complex structures, however, was exclusively Soviet, for example, in the factory instructions for assembling / disassembling and repairing the first keyboard adding machines, the list of instruments officially included a hammer, heavy by the standards of computer technology (even the weight is indicated separately - "150-200 grams" ) and files for fitting parts in place. The album of schemes "Assembly and adjustment of computers VMM-2 and VMP-2", released by KZSM in 1966, describes such wonderful things as:
the turn of the bevel gears should not be noticeable to the eye ... the bracket is set in place, it can be displaced, deflected and straightened ... make a cut when adjusting ... adjust the bracket by filing.
And other high-tech operations.
As you know, much more complex Burroughs machines were produced on the assembly line at the beginning of the XNUMXth century, while in the USSR the assembly of BMM was an individual art, the result of which depended only on the straightness of the craftsman's hands.
Returning to the tabulators, we note that in order to solve the problem with computing technology in 1923, the first new computer technology enterprise in Soviet Russia was created - the Bureau of Precision Mechanics of the 1st Moscow State University, although it does not yet produce, but only repairs the existing adding machines and typewriters. In 1928, ZSPM appeared in its place, on which the production of the first Soviet adding machine "Soyuz-1" (another clone of Odner) began. The Dynamo plant opens in Kharkiv, producing another one-man clone - "Original-Dynamo".
Since the 1930s, ZSPM has been renamed into the First State Plant of Counting and Analytical Machines (the famous CAM, later attached to ITMiVT, which existed until 2010, produced the M-20, BESM-6 and even Elbrus computers - in fact, all machines Lebedev and Burtsev) and, finally, it mastered the production of more complex structures - tabulators CAM T-1 (1935, only allowed to summarize and print the result) and T-2, simplified clones of IBM.
The first domestic balancing tabulator (performing both addition and subtraction) was designed in 1938 under the leadership of V.I. Ryazankin and was produced as SAM T-4 from 1939. At the beginning of 1941 N.I. Bessonov suggested using an electric pulse counter (instead of an electromechanical one) in the T-4 model, which greatly simplified the design. In the mid-30s, the standard binding of tabulators was copied - an electromechanical puncher, a sorting attachment, and so on.
Tabulators in the USSR were dearly loved and counted on them until the 1970s, and on models that did not differ much from machines half a century ago.
In 1948, the S45-1 and C80-1 sorter appeared, in 1950 the balancing tabulator T-5 was produced (I.A.Rakhlin, I.S. Evdokimov). In the 50s, an electromechanical puncher P80-2 was created with automatic feeding and depositing of cards and with a duplication mechanism that made it possible to make punches from previously punched cards. Tabulators T-5M, T-5MU and T-5MV worked on the so-called. machine counting stations for many years. The first MSS was created by the Kharkov Institute of Labor in 1925 and was used for scientific calculations, and in 1928 the employees of the KhIT organized the first station at the Hammer and Sickle machine-building plant.
The profession of "MCC operator" existed until the early 1980s - and this is in the era of the triumph of personal computers!
ITMiVT
And here we smoothly come to the history of ITMiVT, since it was originally organized in 1948 to calculate (mechanically and manually!) Ballistic tables and perform other calculations for the Ministry of Defense. Its first director was Lieutenant General N.G. Bruevich, a mechanic by profession, under him the institute was focused on the development of electromechanical differential analyzers, since the director did not represent any other technique.
In mid-1950, Bruevich was replaced by M.A. Lavrentiev. The displacement came out through a promise to the leader as soon as possible to create a machine for calculating nuclear weapons... And only then Lebedev appeared in ITMiVT.
A similar ITMiVT picture was observed even earlier - in the 30s and 40s, according to the memoirs of Academician Lavrent'ev, back in 1935 at the V.I. Steklov (the largest mathematical center in the country), a computing laboratory was created that carried out orders for large calculations. Moreover, it did not even have tabulators - first-class mathematicians, armed with a pencil and an eraser, played the role of machines, for hours calculating the parameters of steel melting, hydroelectric dams and other trivial things. The orders became more and more, by 1947 the tabulators had not been delivered, and the laboratory of 2 rooms spread over an entire floor, occupying more than half of the institute's area. Recall that it was not the janitors who were engaged in ordinary calculations, but scientists, whose use instead of calculators was even worse than hammering nails with a microscope.
Information about the first digital computers came to the USSR from various foreign sources, of course, intelligence (given how many American technical secrets were stolen in the forties, including the atomic bomb itself, this is not surprising).
Even Malinovsky writes:
Perhaps, to the final decision to engage in the development of a digital computer S.A. Lebedev was pushed by M.A. Lavrentiev. This opinion was expressed by V.M. Glushkov, M.G. Kerin (who together with S.A. Avramenko programmed the first problem for MESM) and O.A. Bogomolets. The latter in 1946-1948, carrying out government orders [that is, spying, neutral Switzerland throughout the Cold War was a vestibule between communism and capitalism and one of the main arenas of the battle of intelligence. - Approx. author], visited Switzerland several times. An avid radio amateur, he collected brochures and magazines of interest to him about digital computing devices. Arriving in Kiev in the summer of 1948, he showed the magazines to Lavrentiev, the latter to Lebedev.
An employee of SKB-245 professor A.V. Shileiko also recalled:
Probably the story with the adoption of I.V. Stalin's decision to develop VT in the country was not so simple. There was a letter to him from Academician M.A. Lavrent'ev, I also assume that intelligence information about K. Zuse's Z-4 vehicle, which was acquired by the Institute of Technology in Zurich, had already passed through Switzerland at that time. There was a report about a rather strange business trip there by a well-known Soviet scientist.
Here we are talking just about the trip of Bogomolets, however, Z4 was never secret, and its description appeared in the journal Mathematical Tables and Other Aids to Computation in an article by Lyndon RC The Zuse computer back in No. 20, October 1947.
Naturally, ENIAC also made a fuss, and immediately after the appearance of the first press releases, on April 5, 1946, the Moore School of the University of Pennsylvania received a letter from A.P. Malyshev with a request to consider the possibility of manufacturing according to the Soviet order "Robot-computer "(Robot Calculator). The dean of the Moore School, Harold Pender, asked for permission from the military, and, judging by the tone of his letter, he really expected to receive it, but this did not happen.
In the fundamental three-volume book “The Atomic Project of the USSR. Documents and materials: In 3 volumes. " (ed. LD Ryabeva. M .: FIZMATLIT, 1998-2009) we find a description of total espionage not only in the field of military technology, but also in computing. A.S. Feklisov writes:
In 1942-1946, I had seven agents in touch - sources of important classified materials. These were agents from among the engineering and technical personnel who held managerial positions in factories and laboratories of RCA, Western Electric, General Electric, Westinghouse, as well as two leading aviation companies, where US military equipment was produced. Especially valuable agents worked in electronics, they gave us more than twenty thousand pages of secret documentary scientific and technical information, including on the latest at that time various types of radars, sonars, sighting systems, anti-aircraft radio fuses, computers and many other devices. Were obtained classified materials on the technology of production and samples of klystrons, magnetrons and other vacuum devices.
Open discussions were also enough.
Since January 1946, Academician N.G. Bruevich seminar on precision mechanics and computer technology, which was attended by employees of the Department of Precise Mechanics of the Institute of Mechanical Engineering of the USSR Academy of Sciences, the Department of Approximate Computations of the Institute of Mathematics. V.A. Steklov Academy of Sciences of the USSR, some laboratories of the Energy Institute named after G.M. Krzhizhanovsky Academy of Sciences of the USSR, as well as several organizations from other cities.
The purpose of the seminar was to exchange the research results of the departments, which later formed the ITMiVT team. Apart from Bruevich, the participants of the seminar were L.A. Lyusternik, I. Ya. Akushsky, M.L. Bykhovsky, I.S. Brook, L.I. Gutenmacher and others, and the materials were published in "Uspekhi Mathematical Sciences" and in "Izvestia of the Academy of Sciences of the USSR".
It was at this seminar that speeches dedicated to computers were made for the first time: in 1947 M.L. Bykhovsky, who at that time was one of the main translators of foreign literature on computers, reported on the Harvard Mk I, and in 1948 he also translated and published an article for Douglas Rayner Hartree's UMN "The Eniac, an Electronic Computing Machine ", Published already in the popular science Nature in 1946. He also made a compilation of the above information and for the first time in the USSR outlined the basic principles of building digital computers in the article" Fundamentals of electronic mathematical machines for discrete counting "(Bykhovsky ML // Uspekhi Mat. 1949 Vol. 4. Issue 3).
At the center of all this was Lavrentyev - a real fanatic and a locomotive for the introduction of digital electronic computers. He tirelessly gathered around him all who were related to machine computing, pushed the development of computers in all ways. And even with the help of a letter to Stalin, he fled Bruevich from the post of director of ITMiVT, when it became clear that there would be no sense from the old man.
Bruevich, in all possible ways for him, tried to direct the efforts of scientific workers to create computers of continuous operation, which objectively delayed the creation of electronic digital machines, Lavrentiev wrote later.
In 1947, he spoke at the jubilee session of the Department of Physical Sciences of the USSR Academy of Sciences:
... I will focus on one more area, the development of which is especially important for applications of mathematics. I mean machine mathematics. If in the main sections of mathematics by the 30th anniversary of the Great October Socialist Revolution we can report: we have caught up, and in many sections we have surpassed foreign mathematics, then with regard to machine mathematics we still need a lot of effort to solve this problem ... Computing cell created in 1935 at the Mathematical Institute. V.A. Steklov, begins to perform, especially in recent years, large-scale calculations. Over the course of 12 years, this cell of two rooms has expanded to an entire floor and now occupies more than half of the entire area of the Mathematical Institute. Further, the Department of Approximate Methods has nowhere to spread at the Institute, in addition, its tasks are such that their solution requires a completely different scope. I would like to say that the decision of the FMSO on the creation of a special Institute, made more than two years ago, will find an early and complete resolution.
The famous oak-headed minister Parshin, whose statements we have already cited, was categorically against the development of computers and said to Lavrentiev:
... when I had to solve a problem, I took 500 students, seated them, gave each of them formulas, and everything was done in two days. And you say - machines!
As a result, the USSR mastered to detonate a nuclear bomb in 1949 without using a computer (like the Americans before), but it became clear that for the further development of atomic weapons (and especially thermonuclear), computer technology is more powerful than "Felix", tabulators and 500 students with pencils , vital.
At the insistence of the Bureau of the OTN of the USSR Academy of Sciences, on September 2, 1949, an order was issued at ITMiVT to create a group to carry out preliminary work on the development of computers. The interim leadership of the group was entrusted to Bykhovsky.
At this time, ITMiVT did not yet know about the work of S.A. Lebedev in Kiev. Only in January 1950 M.A. Lavrentyev and S.A. Lebedev first visited ITMiVT, where N.G. Bruevich introduced them to the work of the institute and to its employees.
Summing up the above, we can conclude that information about the creation of computers in the West, both relay and lamp, was absolutely available to Soviet engineers both through open channels and through closed ones since 1946, as well as information about the architecture of such machines. Because of this, it is difficult to regard Lebedev as some particularly brilliant visionary who created a computing system that had no analogues in the world. He, of course, was a talented engineer and organizer, he was well versed in computer technology of the 1940s and even in the conditions of Kiev dilapidated by the war, selecting from thousands of lamps manually not defective, he was able ... well, in general, to repeat the exploits of British engineers, Tom Kilburn (Tom Kilburn, Frederic C. Williams, Maurice Wilkes, Alan Mathison Turing, Ted Cooke-Yarborough, who were building Manchester Mark 1, EDSAC, Pilot at the same time ACE, Elliott 152 and Harwell Dekatron Computer.
So our MESM clearly does not fit the description of "the main road of the world computer engineering for several decades ahead".
Let's move on to Moscow, to BESM.
According to the memoirs of P.P. Golovistikov, one of Lebedev's associates, held discussions at ITMiVT seminars, during which skepticism about digital technology was shown:
First of all, there was great doubt about the reliability of computers containing several thousand vacuum tubes (the average lamp life was 500 hours and the individual reliability of each lamp was not guaranteed) ... At the same time, any failure in a thousand-lamp computer would lead to an incorrect result. Such bulky, complex and expensive machines will be ineffective or ineffective at best.
The second point of doubt was caused by the complexity and cumbersomeness of the process of preparing a problem for computing on a computer. At that time they did not know anything about mathematical software ... If the solution of the problem itself on a machine takes several minutes, then the preparation of the problem may take several days, or even months. This narrows the range of tasks to be solved. On a computer, it is advisable to perform tasks of the same type that require a very large amount of computation, but a fairly simple program. The user can only be a highly qualified specialist who is well aware of the machine itself, its structure. Therefore, computers cannot find wide application; they can only be used in large computing centers with a large number of mathematicians preparing problems for it. There were other, less substantiated statements about computers. For example: the range of tasks requiring a very large amount of computation is not so large, therefore, computers cannot find widespread use.
The second point of doubt was caused by the complexity and cumbersomeness of the process of preparing a problem for computing on a computer. At that time they did not know anything about mathematical software ... If the solution of the problem itself on a machine takes several minutes, then the preparation of the problem may take several days, or even months. This narrows the range of tasks to be solved. On a computer, it is advisable to perform tasks of the same type that require a very large amount of computation, but a fairly simple program. The user can only be a highly qualified specialist who is well aware of the machine itself, its structure. Therefore, computers cannot find wide application; they can only be used in large computing centers with a large number of mathematicians preparing problems for it. There were other, less substantiated statements about computers. For example: the range of tasks requiring a very large amount of computation is not so large, therefore, computers cannot find widespread use.
Note that Lebedev supported most of these remarks from the very beginning.
He saw computers exclusively as extremely powerful automatic scientific calculators. He had little interest in control machines and was absolutely not interested in economic and general-purpose machines. Hence his passion for programming exclusively in machine codes (or, at worst, autocode), dislike for JLU, exclusively real arithmetic of all his creations (he did not recognize fixed-point arithmetic much more suitable for economic and managerial calculations) and rather specific architectural features, both machines and command systems (we will talk about this in more detail in the part about BESM-6).
Lebedev did not see computers outside of a very narrow specialization - the number of crushers of differential equations. In fact, all his life he approached his dream - to create the most powerful machine of this kind and MESM, BESM-2, BESM-4 were successive iterations of the same architecture, sharpened for one task. Lebedev also never got over his skepticism about two things - integrated circuits and parallel computing. None of his architectures have ever used either one or the other.
In 1950, preparations began for the creation of a thermonuclear bomb. It was no longer possible to do without full-fledged computers.
In the process of research, it turned out that the calculations of the Landau group (for the RDS-6t bomb) could not be completed in time by July 1951, since the methods that are usually used by theoretical physics and that KB-11 and Landau counted on when determining the timing, have been tried but proved to be unusable. As a result, as we know, most of these calculations were performed on Strela.
In general, the appearance of Soviet large machines of the 1950s was determined by the competition between two groups - SKB-245 (MMiP) and ITMiVT (Academy of Sciences of the USSR). Lavrentiev described the whole point in one phrase:
It was not just a development. It was a struggle.
The funny thing is that according to the Decree of the Council of Ministers of the USSR No. 2369 of June 30, 1948 on the creation of the ITMiVT of the Academy of Sciences of the USSR, the joint activities of these departments were prescribed, the MMiP of the USSR was a shareholder in the construction of the institute, it was also supposed to become the customer for the development of new technology. At the same time, MMiP actually disrupted the construction time of the main building of ITMiVT, organized its own design bureau SKB-245 under its wing (which in the 1950s had absolutely phenomenal resources, probably the best in the country, access to all types of memory - from delay lines to prototypes ferrite, later - access to experimental transistors, etc., as well as the relative freedom of creativity) and, taking advantage of the fact that only it allocated resources - allocated everything possible to "Strela" to the detriment of BESM.
By the end of 1955, another vehicle was created in SKB-245 - "Ural" (BI Rameev).
Parshin (the one who at first wanted to emulate a computer with the help of 500 students) sensed huge budgets and grabbed onto them with a stranglehold. In the long term, he generally wanted to kick the Academy of Sciences of the USSR out of the computer business, directly inviting scientists to focus on theory, and give the production of computers to the ministry, which had a research and production base in the form of the Moscow plant of calculating and analytical machines (CAM), SKB-245 based on this plant , as well as Scientific Research Institute Schetmash.
All this was also superimposed on the invisible presence on the horizon of Stalin, who was promised that the computer would be built.
As a result, ITMiVT themselves merged the first round outright, voluntarily supporting the transfer of all resources to Strela, the subtle and cunning politician Academician Keldysh understood that it would not be possible to overcome Parshin anyway. At the same time, the struggle would have exhausted both players, and the project would have failed, as a result, all participants would have gone to build not a computer, but barracks in Siberia.
The leaders of the atomic project generally looked at computers very narrowly, Parshin eventually recognized its usefulness, but considered that 2-3 computers per country would be enough. In this regard, the leadership of the Academy of Sciences went not far from the minister, but believed that computers would also be useful to them, so they advocated expanding their use from the military-industrial complex for general scientific tasks.
What is funny, on the initiative of MMiP, the very fact of the computer's existence was initially strictly classified (as we remember, while sitting in Moscow, the groups of Brook and Lebedev did not even know about each other's work, and SKB-245 was generally classified to death).
And then suddenly in 1953 the secretary general dies. An entire era is ending.
The time is coming for Khrushchev - much more daring, open to innovation and not so fixated on total control, but fixated on “catching up and overtaking”. Lebedev just soars under Khrushchev. By that time, Lavrentyev had already ceded to him the chair of the ITMiVT director, Lebedev was immediately elected an academician and, first of all, was establishing political ties with SKB-245.
The secret of success is simple and described by his colleague Lisovsky:
In general, Sergei Alekseevich treated the party functionaries with trepidation. Everyone understood what a great role the Communist Party played in the country and in the fate of each person individually.
Not being an outright sycophant, Lebedev possessed an amazing feature that determined the fate of the entire Soviet computer school - unlike Kartsev, Yuditsky, Brook, Rameev, Staros - he knew how to be liked. He treated officials with pronounced respect and extreme respect, constantly demonstrated it, never argued with anyone, and in every possible way supported the idea loved by Soviet ministers, brilliantly expressed by Minister Shokin,
You are not creating, the Communist Party is creating.
Lebedev admitted the patriotic role of the CPSU and the ministers personally without the slightest problem, and for this he was exalted.
On the basis of BESM, the Lebedevites are developing two machines at once - completely identical to it, but the serial BESM-2 and a slightly more improved M-20. "Arrow" has been forgotten as a bad dream, both creations of ITMiVT are going into series. The preparation of documentation for BESM traditionally takes 5 years, and it goes into series only in 1958, before 1962 67 pieces were manufactured.
At the same time, the preparation of an optimized version of BESM - M-20 with double speed (20 KIPS) and a reduced number of components - 4 lamps instead of 000 begins. 5 years they managed to assemble only 000 cars, they looked the same: a la "Arrow" and Lebedev's prototypes - cabinets along the walls.
It was M-20 that Kisunko categorically rejected, putting it bluntly:
This craft is not suitable for our needs.
Separately, it was annoying that, despite the similar architecture and the fact that all machines had an extremely cumbersome three-address command system, the sizes and types of words were significantly different. In MESM, there are 20-bit instructions and 17-bit numbers (by the way, another iconic feature of Lebedev's machines - he never matched the sizes of instructions and data, this was brought to the limit in BESM-6 and gave countless hours of joy to programmers). In BESM, there are already 39 bits per number, the command is divided into blocks: the operation code is 6 bits and three address codes 11 each, M-20 was already 45-bit, and so on.
The side branch of the M-20 was the M-40, created by Burtsev, in which the command width was equal to (20? 39? 45?) ... Nope, 36 bits! It seems that the development team was having fun with inventing a machine that, in principle, is not compatible with anything from their previous work.
M-40 was completed by 1960 and, together with its sister M-50, successfully worked in the first launch of an anti-missile, so Burtsev (more precisely, his boss Lebedev) and ITMiVT entered the pantheon of missile defense builders for the first time.
M-20 was assembled 63 at the Kazan plant and it is not known how many at the SAM plant (only 70-75 pieces). It was used in air defense, but it also fell to scientists a little, mainly to nuclear scientists.
The development of the line was the transistor M-220, developed within the walls of ITMiVT without the participation of Lebedev and Burtsev V. Gurov, N. Egorycheva, G.G. Zotkin, V.S. Klepinin and A.A. Shulgin. In fact, it was the same M-20, but converted according to the BESM-6 type for transistors. The alteration was completed in 1968 and until 1978 it was stamped with a record circulation by the standards of the USSR: M-220, M-220A and M-220M - more than 260 pieces, M-222 - 551 pieces.
Struck by the monstrous, by the standards of the mid-70s, the squalor of its circuitry. Diode-transistor logic on fossil transistors P-401, wired on boards 200x120 mm. The EU series, which by that time was already walking around the country with might and main, used GIS, as well as the machines of Kartsev and Yuditsky, in the West, by the mid-1970s, people were gradually moving from integrated circuits to microprocessors, ITMiVT was stuck in the days of the youth of its great boss, when the transistor was considered rocket science.
It is not surprising that with such an element base, only 220 KIPS were squeezed out of the M-27 - a penny at that time. What is absolutely amazing is that with a complete change of the element base from lamps to transistors, it would seem that the speed should rise by an order of magnitude, as it always happened. However, the M-20 produced 20 KIPS, and the M-220 only 27 KIPS, and not 100-150, as the minimum would be expected from a machine assembled on elements 10 times faster. What was the matter there - it was not possible to figure it out, the secret of assembling machines of the 2nd generation, which barely surpassed the 1st in speed, was available only to ITMiVT engineers and was lost for centuries.
What's even more piquant is that neither the size (!) Nor the power consumption (!) Of this miracle of technology has changed much. M-220 required mash. hall of 100 sq. meters (for a performance of 27 KIPS!) and ate 20 kW (not counting cooling), the M-20 required a hall of 170 meters and 50 kW. Again, when switching to transistors, these values usually changed much more significantly, for example, the IBM 7000 transistor line increased performance tenfold compared to the IBM 700 tube, reducing the size and power consumption by more than three times.
The evolution of a computer through the eyes of a healthy person - a standard IBM 709 module and a completely similar module IBM 7090
The command system was also Lebedev's classic - that is, a bulky three-address, a magnetic drum, no less fossil than transistors, was used as a buffer memory! In those years, it was already archaic, comparable to harnessing a horse to a cart compared to a car.
No bourgeois excesses were envisioned in the original version of the machine - in a world where workplaces with terminals have become the norm since the mid-1960s, the M-220 was controlled according to the classics - from the remote control and gave the result of work in a harsh Soviet way, printing it out using an ADCP- 128 or punch.
The M-222 received an incredible upgrade of the ergonomics of the human-machine interfaces: it was not a custom control panel with a bunch of buttons that was adapted for data entry, but a standard Czech typewriter Consul-254 (in fairness - different versions of Consuls - 254, 256, 260, 260.1, 260.2 were massively used not only in the M-220 line, but also in almost all Soviet mainframes, the magic of CRT monitors as a terminal was not yet discovered in the Union at that time).
A deep modernization of the M-220 - M-222 was overclocked in 1971 to 40 KIPS (which is surprising - the same amount was produced by the M-40 Burtseva tube back in 1960).
M-220 was used in some universities, for example, MGTU im. Bauman, but they were mainly made for the needs of the Ministry of Defense, the computer was part of the URTS-2M telemetry complex.
Let's note a funny moment - the M line of ITMiVT received performance indices in KIPS, again, the tradition of Lebedev, who called the M-20 so, being proud of its power (10 times higher than Strela, 10 times less than its modern transistor IBM 7030) , the initiative was supported by Burtsev with the M-40 and M-50, according to the logic of things, the M-220 was supposed to produce 220 KIPS, but something went wrong.
Kartsev, by the way, calling his monster modestly, M-9, thus harshly spiked ITMiVT and Ded personally, saying at a conference in the Novosibirsk branch of the USSR Academy of Sciences:
We named our machine M-9, since its speed is 10 ^ 9 degrees of operations.
Let's return to Lebedev and his creations.
As we have already said, after Stalin's death, things went uphill, the BESM, which had not even got a normal RAM, in 1955 was studied at the commission for consideration of applications for computing work for the Computing Center of the USSR Academy of Sciences, formed in the same year. The commission was made up of Lebedev's old friends - academicians M.A. Lavrentieva, L.A. Artsimovich, A.A. Dorodnitsyn and M.V. Keldysh. The summary was simple, Keldysh said as he snapped:
I believe that the calculations carried out on the BESM machine by the Department of Applied Mathematics of the MI of the USSR Academy of Sciences are of paramount importance and are more important than most of the calculations currently carried out by other organizations in the Computing Center ... hours of machine time per week.
As a result, BESM received a normal memory and entered the atomic project.
Why did Lebedev, without waiting for a positive decision on the BESM line, start a parallel project - the M line, in fact, clones of his own architecture?
In the article "The Soviet atomic project and the formation of domestic computer technology" in the materials of the international conference Sorucom 2017, the following answer was proposed:
The very idea of duplicating projects for creating a computer came, perhaps, from the Atomic Project (or, rather, from the department of L.P. Beria), where the institution of understudies, both performers and organizations working on the same topic, was adopted. It is known that when creating the first atomic bombs, two groups of mathematicians and physicists were involved in the creation of the first atomic bombs for the calculation and theoretical substantiation (L.D. Arzamas-70, and from the memoirs of Academician V.S. Vladimirova, it follows that M.A. Lavrentyev was I.V. Kurchatov.
However, this version sounds rather strange - in all cases the doubles worked in parallel and independently, and in the case of Lebedev, he began to duplicate himself in a cunning way.
As a result, both rulers took off, the M-20/220 and the BESM.
On the basis of M-20, but already on semiconductors, a team of graduate students of ITMiVT assembled in 1964 their alternative version of the M-220, a model called BESM-3M. He remained in a single copy and did not represent anything remarkable, therefore there is practically no information about him.
However, on its basis, a serial transistor BESM-4 was created, in the amount of 30 pieces, it was produced since 1965. The performance was extremely dull, at the level of the same antediluvian M-20. The architecture has not fundamentally changed - the same three-address command system, real arithmetic. The chief designer was O.P. Vasiliev (formal, as the head of the SLE, was actually done by the same graduate students).
For BESM-4, there were at least 3 different compilers from the Algol-60 language, a Fortran compiler, at least 2 different assemblers (Dubninsky and Bayakovsky), and a compiler from the original Epsilon language. In general, in the USSR there was a very interesting dichotomy - practically no chief designer of something designed this something, but was just a boss. On this occasion, Lebedev himself spoke quite ironically, according to the memoirs of Malinovich:
I remember that in a conversation about defending his dissertations, he remarked, not without irony, about his institute: "And we have a division of labor: some make machines, others defend themselves."
Let's note one more funny moment.
In November 1953, a third independent computer appeared in the USSR, created by the staff of the Institute of Atomic Energy for themselves. She had a TsEM-1 index and worked there for 7 years. It all started in a classic way: deputy. Kurchatov, Academician Sobolev read about ENIAC and showed the article to young specialists, led by G.A. Mikhailov. He got the idea in the same way, dug up the EDSAC schemes among the Western press and four of his friends cloned this machine.
It turned out to be simple and slow - 1 lamps, RAM for 900 binary 128-bit numbers on mercury delay lines of 31 numbers each, with sequential sampling at a frequency of 16 kbit / s. The memory capacity was later increased to 512 numbers and an external memory was added, 496 numbers on a magnetic drum. Data input and output was organized on the basis of the ST-4 telegraph apparatus. The performance is about 096 KIPS. The machine was housed in 35 racks and consumed 0.3 kW.
From the amusing details, we note that the modes in the main blocks could be monitored on an oscilloscope, at least some kind of display.
In the process of installing the machine, Lebedev himself visited the institute, then a scene followed, which we will allow Mikhailov to describe:
Having visited our laboratory and meticulously examined TsEM-1, Sergei Alekseevich surprised us with the question: "Don't you knock on it with a sledgehammer?" It turned out that on BESM a sledgehammer is a standard tool, and its blows on the iron frame of the machine is one of the elements of prevention! Equally surprising would now seem to be an order not to allow the solution of the problem for more than 15 minutes without recalculation so as not to waste machine time.
Lebedev, in addition, rejected the unicast scheme (as we remember, he recognized only hardcore, only three-address architecture in the spirit of the 1940s), as a result, Mikhailov partially caved in before the great authority and converted the car into a two-address one, for which he had to disassemble some of the ready-made racks ...
As expected, the novelty was first subjected to obstruction within the walls of his native institute, academician Lev Andreevich Artsimovich did not recognize the value of the machine - Mikhailov obtained a solution to the plasma compression equation on it, which contradicted all his theoretical calculations. Later he had to change his mind, the result of the experiment was extracted from the archive, confirming the correctness of the calculations (initially he was rejected as clearly erroneous, because he did not agree with Artsimovich's theory).
What conclusions can we draw from the above?
Lebedev was a talented designer of the 1940s, however, he did not stand out in any way from a number of people like Zuse and Kilburn and was definitely inferior to von Neumann or Turing. He developed the architecture of the BESM / M-20 by the standards of the early 1950s at a decent European level.
Since the mid-1950s, he no longer promoted, but rather slowed down the development of computing technology due to a fanatical adherence to outdated architecture - three-address commands, monoprocessor, strictly transistor machines. He had nothing to do with missile defense and did not create anything as striking and original as the work of Kartsev and Yuditsky.
By the will of fate and his character, it turned out that he suited the Soviet partocrats ideally, like a precision-fit gear and brilliantly played his role, having received a huge number of awards and recognition.
ITMiVT actually turned into the main and only center for the development of computer technology in the USSR, the introduction of ES computers did not stop it in any way - the same BESM-6 were riveted in hundreds almost until the beginning of the 1990s.
In addition, ITMiVT, or rather his school, indirectly survived to this day - from Lebedev to Burtsev with "Elbrus" and from there to MCST and microprocessors.
So we can say that from the point of view of evolution, Lebedev and the clan of his students turned out to be incredibly successful not only in the Union, but also in Russia, having successfully survived the collapse of the USSR. In any case, it deserves respect.
The complete genealogy of ITMiVT machines and their descendants, we will briefly go through this scheme in the next part
But what about the main Soviet computer, the great and terrible, BESM-6, which had no equal in the world and the CDC6600 nervously smoked on the sidelines?
And no less great "Elbrus"?
Our next story will be devoted to the mythology and technoarcheology of these machines, after which we will safely move on to describing the end of the Soviet missile defense program.
Information