Alternative architecture began when Isaac Brook and Bashir Rameev met in early 1947, who were united by a common interest in creating an analogue of ENIAC. According to one legend, Rameev learned about the computer while listening to BBC radio, according to another version - Brook, being connected with the military, knew that the Americans had built a machine for calculating firing tables from some secret sources.
The truth is a little more prosaic: back in 1946, an open article about ENIAC was published in the journal Nature, and the whole scientific world knew about it, even a little interested in computing. In the USSR, this journal was read by leading scientists. And already in the second issue of "Uspekhi Mathematical Sciences" in 1947, a 3-page article by M. L. Bykhovsky "New American calculating and analytical machines" was published.
Bashir Iskandarovich Rameev himself was a man of a difficult fate. His father was repressed in 1938 and died in prison (interestingly, the same fate awaited the father of the second M-1 designer, Matyukhin). The son of the "enemy of the people" was kicked out of the MEI, for two years he was unemployed barely making ends meet. Until he got a job in 1940 as a technician at the Central Research Institute of Communications, thanks to his penchant for radio amateurism and invention. In 1941 he volunteered for the front. He went through all of Ukraine, survived everywhere, atoned for the crime of being a relative of an enemy of the people with blood.
And in 1944 he was sent to VNII-108 (radar methods, founded by the famous engineer - Rear Admiral and Academician A.I. Berg, who was also repressed in 1937 and miraculously survived). There Rameev learned about ENIAC and got the idea to create the same one.
Under Berg's patronage, he turned to the head of the ENIN electrical systems laboratory, Isaac Semenovich Brook.
Brook was a keen electrical engineer, but a minor inventor. But a talented and most importantly - a punchy organizer, which was almost more important in the USSR. For the previous 10 years, he was mainly engaged in taking part, leading and supervising (moreover, he took off to leadership positions immediately after graduating from the institute and subsequently systematically and successfully forged his career), until the creation of a device popular in those years at ENIN, a great analog integrator for solving systems of differential equations. As the project manager, it was Brook who presented him at the Presidium of the USSR Academy of Sciences. The academicians were impressed by the epic nature of the device (an area of as much as 60 square meters) and immediately elected him a member correspondent (though this, however, his career reached its peak, he never became a full academician, despite all his aspirations).
Hearing that computers are being built at ENIN, Rameev came there to present his ideas to Brook.
Brook was a savvy and experienced person. And immediately he did the most important thing in the design of the Soviet computer - in 1948 he filed an application with the Patent Bureau of the State Committee of the Council of Ministers of the USSR for a whole copyright certificate (to which, incidentally, Rameeva also wrote) for “Invention of a digital electronic machine”. Of course, now it looks pretty funny (well, wow, the USSR issued a patent for the invention of a computer, after all the ABC, Harvard Mark-1, Z-1, EDSAC, ENIAC, Colossus and others). But this patent, firstly, allowed Brook to immediately enter the pantheon of Soviet computer creators, and secondly, ranks and awards were relied on for each invention.
The construction of a computer, however, did not work out. Because immediately after receiving the patent, Rameyev was somehow dragged into the army again. Apparently to serve what he did not complete in 1944. He was sent to the Far East, but (it is not known whether Brook intervened or not) a few months later, at the personal request of the USSR Minister of Mechanical Engineering and Instrumentation, P.I. Parshin, as a valuable specialist, sent back to Moscow.
In general, the relationship between Brook and Rameev is full of fog. Upon his return, for some reason, he did not join the M-1 project, but preferred to leave Brook for another party "designer" - Bazilevsky, in SKB-245, where he later worked on "Strela", which competed with Lebedev's BESM (we will cover in more detail this titanomachy in the next issue).
Lebedev lost then. But I didn't go to the second round. And in accordance with the principle “if you cannot win - lead”, he himself started designing the M-20 machine in SKB-245 together with Rameev. In addition, Rameev is known as the general designer and author of the legendary Ural series - small tube machines, very popular in the USSR and the most massive in the first generation.
Rameev's last contribution to the development of domestic technology was his proposal not to use the IBM S / 360 model as an illegal copy model, but instead it is already quite legal to start developing, together with the British, a line of computers based on ICL System 4 (the English version of RCA Spectra 70, which was compatible with the same S / 360). It would most likely be a much better deal. But, alas, the decision was not made in favor of Rameev's project.
We will be back in 1950.
Frustrated, Brook sent a request to the personnel department of the Moscow Power Engineering Institute. And the creators of M-1, about 10 people, began to appear in his laboratory. And what kind of people they were! Not many had completed higher education by that time, some were graduates of technical schools, but their genius shone like the Kremlin stars.
Nikolai Yakovlevich Matyukhin became the general designer, with a fate almost identical to that of Rameev. Exactly the same son of a repressed enemy of the people (in 1939 Matyukhin's father received a relatively humane 8 years, but in 1941 Stalin ordered the execution of all political prisoners during the retreat, and Yakov Matyukhin was shot in the Oryol prison). Fond of electronics and radio engineering, also expelled from everywhere (including the family of the enemy of the people was evicted from Moscow). Nevertheless, he was able to finish school in 1944 and enter the MPEI. He did not get a postgraduate study (again, he was rejected as politically unreliable, despite already two copyright certificates for inventions received during his studies).
But Brooke noticed the talent. And he was able to drag Matyukhin to ENIN for the implementation of the M-1 project. Matyukhin has proven himself very well. And later he worked on the continuation of the line - machines M-2 (prototype) and M-3 (produced in a limited series). And since 1957, he became the chief designer of the NIIAA of the Ministry of Radio Industry and worked on the creation of the Tetiva air defense control system (1960, an analogue of the American SAGE), the first serial semiconductor domestic computer, with microprogram control, Harvard architecture and boot from ROM. It is also interesting that she (the first in the USSR) used forward, not reverse encoding.
The second star was M. A. Kartsev. But this is a man of such magnitude (who had a hand directly in many of the USSR's military developments and played a huge role in the creation of missile defense) that he deserves a separate discussion.
Among the developers was a girl - Tamara Minovna Aleksandridi, the architect of RAM M-1.
General view and arithmetic device M-1 (Journal "Modern Automation Technologies" 2/2012, article by Yu. Rogachev "The first automatic digital computer M-1")
The work (as in the case of Lebedev) took about two years. And already in January 1952 (less than a month after the commissioning of the MESM), the practical operation of the M-1 began.
The paranoid Soviet craving for secrecy led to the fact that both groups - Lebedev and Brook - did not even hear about each other. And only some time after the delivery of the cars did they find out about the existence of a competitor.
Note that the situation with lamps in those years in Moscow was even worse than in Ukraine. And partly for this reason, partly out of a desire to reduce the power consumption and dimensions of the machine, the M-1 digital computer was not purely a lamp. M-1 triggers were assembled on 6N8S double triodes, valves on 6Zh4 pentodes, but all the main logic was semiconductor - on copper-oxide rectifiers. A separate riddle is also associated with these rectifiers (and riddles in stories domestic computers just heaps!).
In Germany, similar devices were called Kupferoxydul-Gleichrichter and were available to Soviet specialists to study captured radio equipment among the mountains. Hence, by the way, the most frequent jargon, albeit incorrect, naming of such devices in the domestic literature as cuprox rectifiers, which suggests that we got to know them thanks to the Germans, although there are also some mysteries here.
The copper-oxide rectifier was invented in the USA by Westinghouse Electric in 1927. Produced in England. From there he went to Europe. In our country, it seems, a similar design was developed in 1935 at the Nizhny Novgorod radio laboratory. Only here there are two but.
Firstly, the only source that tells us about this is, to put it mildly, biased. This is VG Borisov's brochure "Young radio amateur" (issue 100), published already in 1951. Secondly, these domestic rectifiers were used for the first time in the first domestic multimeter TG-1, the production of which began only in 1947. So, with a considerable degree of probability, it can be stated that the technology of copper-sour rectifiers was borrowed by the USSR in Germany after the war. Well, or individual developments were undertaken before it, but it obviously went into production only after studying captured German radio equipment and, most likely, was cloned from Siemens SIRUTOR rectifiers.
Those same cuprocks stumbling blocks in the German radio museum (oldradio.de)
What kind of rectifiers were used in the M-1?
Without exception, all sources speak of the Soviet KVMP-2, this conversation is based on the memoirs of the participants in the events. So, in the memoirs of Matyukhin it is said:
The search for ways to reduce the number of radio tubes in the car led to an attempt to use the KVMP-2-7 cuprox rectifiers, which turned out to be in the laboratory warehouse among the trophy property.
It is not very clear how Soviet rectifiers (especially, the appearance of the KVMP-2 series - this is absolutely definitely not earlier than 1950) ended up among the captured German property a year before their creation? But let's say that there was a slight dip in time. And they got there. However, the developer of the M-1 I / O device, A.B. Zalkind, writes in his memoirs:
From the composition of captured radio components, I.S.Bruk suggested using selenium cuprox columns for signal decoding, consisting of five tablets and connected in series inside a plastic tube with a diameter of only 4 mm and a length of 35 mm.
Leaving aside the mixing of selenium and cuprox columns together (and these are different things), the description shows that the original rectifiers do not correspond to KVMP-2-7 either in size or in the number of tablets. Hence the conclusion - the memoirs in our time cannot be trusted. Perhaps, trophy cuproxes were used on the first models, and when the possibility of their use was proved, then, as the same N. Ya.Matyukhin further writes,
Brook agreed to make a special version of such a rectifier the size of a conventional resistance, and we created a set of typical circuits.
Do you think this is the end of the riddles?
In the description of the next machine M-2, the parameters of the KVMP-2-7 are given, and they are as follows. Allowable forward current 4 mA, forward resistance 3–5 kOhm, allowable reverse voltage 120 V, reverse resistance 0,5–2 MΩ. This data spread all over the network.
Meanwhile, they seem absolutely fantastic for such a small rectifier. And all the official reference books give completely different numbers: direct current 0,08–0,8 mA (depending on the number of tablets) and so on. Reference books have more faith, but how then could Brook's KVMP work if, with such parameters, they would instantly burn out?
And Lebedev was far from being a fool. And he was very good at electronics, including trophy ones. Nevertheless, the idea of using copper-sour rectifiers for some reason did not come to him, although he was a virtuoso in assembling computers from non-standard materials. As you can see, Soviet technoarcheology holds no less mysteries than the tomb of Tutankhamun. And it is not easy to understand them, even with memoirs and memoirs of eyewitnesses of the events on hand.
The text of the first program, which was performed on the M-1 (B. N. Malinovsky "History of computer technology in persons")
In any case, M-1 started working (but even establishing exactly when exactly is an unrealistic task; in various documents and memoirs, the date range appears from December 1950 to December 1951).
It was smaller than MESM and consumed less energy (4 sq. M and 8 kW versus 60 sq. M and 25 kW). But it was also relatively slower - about 25 ops / sec over 25 bit words, versus 50 ops / sec over 17 bit MESM words.
Outwardly, the M-1 looked more like a computer than an MESM (it looked like a huge number of cabinets with floor-to-ceiling lamps all along the walls in several rooms).
We also note that the monstrous battles about who was the first: Lebedev with the Ukrainian group or Brook with the Moscow one, do not subside to this day.
So, for example, despite the fact that the first launch of MESM was documented on November 6, 1950 (which is confirmed by numerous interviews with all the developers, and Lebedev's papers), in the article “History worth rewriting: where the first Soviet computer was actually made "(Boris Kaufman, RIA News) we meet the following passage:
“The fundamental difference between a computer and a calculator is that ordinary differential equations can be calculated on a programmable calculator, but not partial differential equations. The purpose of her [MESM-1] work was to speed up the counting, it was not a universal computing machine for scientific calculations - there were not enough resources to work with matrices, insufficient memory (31 variables) and small bit width, only four significant digits in the decimal system. It is no coincidence that the first production calculations at the MESM were carried out only in May 1952, when a magnetic drum was connected, which made it possible to store and read data, ”writes the Russian historian of computer technology, leading researcher at the Institute of Information Technology of the Russian Academy of Sciences Sergey Prokhorov. But in the M-1, the memory on cathode-ray tubes was initially integrated, and the tubes were taken from a conventional oscilloscope. It was improved by a student at the MPEI Tamara Aleksandridi ... An elegant solution that a young girl found was much better than all foreign computers of that time (all two). They used the so-called potentioscopes, which were developed specifically for the construction of computer storage devices and were at that time expensive and inaccessible.
It is rather difficult to comment on this.
Especially the unique author's definition of a computer and a calculator, which until then has not been found anywhere in a hundred years of development of computing technology. No less surprising is the "unique" superiority of tubes from oscilloscopes as RAM over Williams-Kilburn tubes (as they are correctly called, apparently, in the West they did not know that it was possible to assemble a computer from a trophy radio junk, and for some reason they made expensive and stupid solutions) , as well as the mention of only two (instead of at least 5-6) Western cars of that time.
According to Zalkind's memoirs, one of the first great scientists who showed interest in M-1 was Academician Sergei Sobolev. His collaboration with the creators of the next model M-2 was prevented by an episode in the elections to full members of the Academy of Sciences of the USSR.
Lebedev and Brook claimed one place. The decisive factor was the voice of Sobolev, given by him for his student Lebedev.
After that, Brook (who remained only a member of correspondent for life) refused to provide Moscow State University, where Sobolev worked, with the M-2 car.
And a big scandal erupted, which ended with the independent development of the Setun machine within the walls of Moscow State University. Moreover, its mass production ran into obstacles already from the Lebedev group, who wanted to achieve as many resources as possible for their new M-20 project.
We will talk about Lebedev's adventures in Moscow and the development of BESM next time.
To be continued ...