27

In the USSR there flourished some very interesting machines, including the БЭСМ and МЭСМ lines, the Сетунь, the ЭВМ Стрела and others.

Maybe the most famous ones are

  • БЭСМ-4, which is said to have done the first computer animation,
  • the БЭСМ-6, a 48-bit mainframe supporting the Soyuz project sending stuff to outer space,
  • or the ternary (not binary!) Setun.

These here computers I've listed look like real successes to me. But then suddenly, around 1980 I would guess, the БЭСМ line stopped being developed any further, the 1801s had their microprograms altered to interpret PDP-11 instructions (this presumably means eschewing the native instruction set they had before, Elektronica NC, about which there is very little information online), chips like the КР1858ВМ1 and other Z80 and 8080 clones started being more common, the Toshiba-Kongsberg scandal, etc etc,

I'm not sure how much of this is to do with miniaturization. In the West there are examples of big machines being later implemented on single chips (Harris 6120, TI9900, etc. etc), but I don't see that kind of thing happen with any of the indigenous Russian designs. So I tentatively suggest that this change was not merely a result of technological advancement, but that interest in Western computers grew, and interest in native computers waned.

Why did this happen?

10
  • 19
    The cost of hardware incompatibility is software shortage.
    – Brian H
    Commented Mar 29, 2022 at 12:59
  • 8
    @OmarL You assume the existence of platform independent programming languages and operating systems. This is an interesting assumption pre-System/360. Commented Mar 29, 2022 at 13:56
  • 4
    As a result of the decision of the Council of Minister of the USSR in the late 1960s to focus on producing "compatible" systems (commonly believed to be able to benefit from smuggling software, but, likely, also to be able to benefit from Western peripherals and specifically media, as opposed to the Soviet/Comecon ones which were of notoriously bad quality and reliability), funding for development of original CPU architectures had stopped.
    – Leo B.
    Commented Mar 29, 2022 at 21:56
  • 2
    @Raffzahn, this intra-Soviet institutions competition is a very important point. The competition is not strictly economic but on who convinces the Minister. Playing on traditional Russian suspicion for indigenous designs (amongst other things), the "clone" side won the argument. As for Comecon, it was in their interest; in fact, East Germans already had considerable success cloning IBM.
    – Zeus
    Commented Mar 30, 2022 at 1:11
  • 4
    ...While there are many accounts of this story, for a broader outlook, I can't recommend enough Red Plenty by Francis Spufford. It explicitly touches on this subject.
    – Zeus
    Commented Mar 30, 2022 at 1:13

6 Answers 6

39

I'm not sure your question is correct. You posit that native Soviet designs were ahead until some point around 1980, when "suddenly" clones of Western designs took over. In fact, Comecon countries started cloning Western machines way earlier. Three examples from the top of my mind, cloning the most successful Western designs of their time:

  • There was the ES EVM (ЕС ЭВМ) series of IBM System/360 / System/370 compatible machines. Work began in 1968, production began in 1972.

  • The Hungarian TPA series of machines started with a clone of the PDP-8 in 1968, and later included clones of the PDP-11 (1977) and the VAX.

  • Production of the SM EVM (СМ ЭВМ), a Soviet PDP-11 clone, started in 1975.

The PDP-8 FAQ lists two more Soviet PDP-8 clones (Electronica-100, Saratov-2), but without giving dates. Similarly, the PDP-11's Wikipedia page has a long list of unlicensed clones from Comecon countries, though I don't know how many of those predate 1980.

It's interesting to compare the BESM-6 (БЭСМ-6) from your question with the ES EVM. Both are mainframes designed in the mid-60s, with production starting in the late 60s and ending (with various modifications and improvements along the way) in the 80s. 355 BESM-6 were made, but - if you include production in other Comecon countries - more than 15000 ES EVM.

In general, the number of computers produced in the East was rather low: Wikipedia cites a source stating that "by 1972 the Comecon countries had produced around 7,500 computers, compared to 120,000 in the rest of the world." Reproducing Western architectures meant that software was (more or less) readily available (and definitely easier to smuggle ;-) ), whereas it had to be rewritten for home grown designs.

10
  • 1
    In that context Leo's questions about various aspects of the BESM-6 are quite interesting. For example languages, seem to follow quite close what IBM delivered for the /360 family. Most easy to spot when looking ah his assembler examples.
    – Raffzahn
    Commented Mar 29, 2022 at 14:07
  • @Raffzahn - Sorry, you've lost me. Which questions? Which assembler examples? Commented Mar 29, 2022 at 14:11
  • 2
    Various questions by Leo/besm-6
    – Raffzahn
    Commented Mar 29, 2022 at 14:28
  • 1
    Yes, you are right, IZOT did start production in 1968, but the IZOT 1016S - which was the PDP-11 clone came later - there might have been a PDP-8 clone in 1968 - but I can't find anything about it Commented Mar 30, 2022 at 17:58
  • 1
    It was the IZOT-301 - PDP-8 clone Commented Mar 30, 2022 at 18:02
19

ETA:

You don’t incorrectly claim that the Soviet Union was ever ahead of the United States in computing, but you use similar examples to some people who do. If so, you should appreciate the accomplishments of Soviet computer engineers in the proper context. It produced some successful machines that were not based on Western ideas, but it was always behind. From your reference to the BESM-6 and the Soyuz mission, you might have heard this myth from someone like Gavin Mendel-Gleason. (It seems to be amplified by modern-day Communists, who want to believe the Soviet Union really was the technological marvel its propaganda claimed.) It seems to trace back to an especially uncritical reading of Boris Malinovsky’s Pioneers of Soviet Computing that goes beyond even the claims the author made. Typically, it says that the BESM-6 was completed in 1965, was the fastest computer in the world at the time, and that during the Apollo-Soyuz joint mission (in 1975!) completed a task thirty times faster than American computers. At least the first two of these claims are incorrect. The third is uncertain (It comes from combining two ambiguous sentences from Malinovsky’s book.) and irrelevant even if accurate.

The BESM-6 completed its design work in 1967 and went into production in 1968. Malinovsky says, “The BESM-6 prototype was tested in 1965; in 1967, the first manufactured model had its trial run.” (I’ll be generous and compare to American machines that were in production before 1965.) It ran up to 1 MIPS, although only for a certain subset of single-address instructions, at a clock speed of 10 MHz. In comparison, the IBM 7030 Stretch, built in 1961, already ran at 1.2 MIPS (although the instruction sets weren’t directly comparable). In 1964, the CDC 6600 supercomputer was two to three times faster than the BESM-6 would be when it was finished in 1967, running at the same 10 MHz clock speed.

By 1975, when the Apollo-Soyuz mission took place, the CDC STAR-100 supercomputer, one of the first to support vector instructions, ran at 100 MFLOPS, one hundred times faster than the BESM-6. NASA just didn’t happen to use one for that particular mission. This allowed the Soviets to claim a propaganda victory, although Boris Malinovsky, the source for this claim, is ambiguous about what he is actually comparing, and gives no citations. Even on a favorable account, the Soviets were continuing to build ten-year-old supercomputer in 1975—Malinovsky actually says these calculations were done by “a new computer complex that included a BESM-6 and other domestic high-speed computers”—over a time period when American computers got thirty-three to fifty times faster than they were in 1965.

The history of Soviet computing is not going to make any sense to someone who mistakenly thinks that the Soviet Union was anywhere close to the United States in computing in the 1970s, much less that it was far ahead.

Original Answer

I highly recommend Jimmy Maher’s series, “Tales of the Mirror World,” posted on the Digital Antiquarian blog. (Which anyone who scrolls this far down the page should definitely be reading!)

In particular, part II covers the inflection point, around 1970, when the Soviet Union gave up on designing systems like the BESM mainframe, which was technically on par with Western computers despite not being made from integrated circuits, but which the Soviet Union was unable to produce in quantity. The main problem it had was not in its computer designs, as the Rand Corporation noted at the time, but its ability to produce them. In 1970, the Soviet Union only had approximately five thousand computers.

The article gives as an example:

A popular story that circulated around Western intelligence circles for years involved Viktor Belenko, a Soviet pilot who went rogue, flying his state-of-the-art MIG-25 fighter jet to a Japanese airport and defecting there in 1976. When American engineers examined his MIG-25, they found a plane that was indeed a technological marvel in many respects, able to fly faster and higher than any Western fighter. Yet its electronics used unreliable vacuum tubes rather than transistors, much less integrated circuits — a crippling disadvantage on the field of battle. The contrast with the West, which had left the era of the vacuum tube behind almost two decades ago, was so extreme that there was some discussion of whether Belenko might be a double agent, his whole defection a Soviet plot to convince the West that they were absurdly far behind in terms of electronics technology. Sadly for the Soviets, the vacuum tubes weren’t the result of any elaborate KGB plot, but rather just a backward electronics industry.

The Ryad project that began in 1970 marked a departure from the BESM-6, in that it was a copy of the IBM 360. It would not be the last: Leo D. Bores, an American surgeon who saw the Agat microcomputer called it a copy of the Apple II, and a bad one. DEC was aware that the Eastern Bloc was copying its designs, and began etching messages in pidgin Russian into their CPUs, where they would show up on X-ray scans. One of them was intended to say, “CVAX: When you care enough to steal the very best.” The East German team reverse-engineering the CPU, either unfamiliar with the Hallmark greeting-card slogan the Americans were riffing off of, or wanting to flatter their Russian overlords, told them that it said, “Why steal? You already have the best.” Similarly, the KP580BM80A is, as the name suggests, a copy of the Intel 8080.

Up until the 1970s, the Soviet Union had sometimes been able to buy chip-fabricating equipment from the West, but the Carter and Reagan administrations put export restrictions on American technology that ended those transactions. Starting in 1970, the Soviets began trying to split up the design and manufacturing of different computer components among its Warsaw Pact allies.

There was at least one locally designed computer of the ’80s that deserves mention, however: the Electronica BK-0010. Its CPU was compatible with the instruction set of the American PDP-11, but was an original design, and the machine was a 16-bit personal computer at a time when most American home computers were still 8-bit. It cost one-fifth as much as the Agat, although that meant there was always a long waiting-list to get one, and some people instead tried putting one together from kits. It was created to fulfill a promise by the new Soviet Premier, Mikhail Gorbachev, to put a computer in every classroom.

1
  • 29
    Sadly, as I write, I just learned that Club 8-bit, one of the most important museums of Soviet computer history, and its more than 500 vintage computers, was destroyed in an attack on Ukraine.
    – Davislor
    Commented Mar 30, 2022 at 2:01
14
  1. Economy of scale.

    The "scale" part requires an economy to back the development and the production by buying the products. The iron curtain separated the "west" (an economy with ~1 billion people) and the "east" (~300 million people with much less GDP per capita). Other parts of the world are not counted because they came late to the party.

  2. Administrative approach vs free market approach. Only the first western projects were government-run (with the usual governmental overhead). Then, the market took off.

  3. A lot of effort in the east went to making the designs intentionally different than the western approach to the same task - and not necessarily better. Here goes Setun as well.

  4. Once sufficiently lagged and constrained in R&D resources, the east was much more interested in cloning and reverse-engineering the western designs. This was much more successful than reinventing a wheel after a wheel.

And then, what exactly did the East better? (I am from the East, I know the answer.)

3
  • 6
    Didn't the best DOS computer viruses come from the east? (Back when these were technological marvels, not steal your money things.)
    – Joshua
    Commented Mar 29, 2022 at 18:37
  • 1
    Yes, there was a brief period starting in 1990 when lots of major state-owned IT companies with very skilled engineers specialized in reverse engineering - a very rare skill in the commercial IT world - became suddenly unemployed. By 1995 most of them were employed in the new private IT sector and the virus writing came to a halt. There was the Dark Avenger from Sofia, one of the best known legends of this era. Commented Mar 30, 2022 at 17:16
  • what exactly did the East better? -- the 3" floppy disk.
    – chx
    Commented Mar 31, 2022 at 0:21
6

There is still development of original computer architectures in Russia, but they don't seem to get used commercially very much. In the post-Soviet era, most businesses preferred imported computers, for both performance and reliability.

The Elbrus brand name covers several ISAs, which are mostly used in space and defence systems.

5

COCOM and similar things were mentioned. But that "marvellous design" OmarL mentioned was seen in the USSR at the time as heterogeneity and inefficient fund spending. They wanted a one size fit all, something similar to the position of the IBM at the time.

In 1969 the decision came to copy Western designs. (It was made by a joint ministry group, as far as I understand.) This decision is widely cited as the beginning of the downfall of the Soviet computer development, and not only by Russians themselves. The Soviets switching to IBM was seen as a great success for US. The scaling-down of the hardware, especially in the late 80s, additionally made the gap larger with the time.

2

Maybe another point not mentioned so far that may help to put this 'overtake' into perspective:

  • The very same happened in the 'West'

The 70s and 80s were a time of many designs losing out against the few flourishing. Germany may be a good example to illustrate this as there were three quite successful manufacturers for large machines in 1960/1970s:

  • Siemens creating the 2002 (*1)
  • Zuse offering their Z22/Z23
  • Telefunken with very successful TR4/TR440 machines (*2)

Each sold between 50 and 100 units, none the less, all were gone by the mid 1970s. Sure, the solution was a 'capitalist' one, with Siemens buying Zuse KG in 1967 and switching IBM compatible designs in 1968 and later, in 1974, buying Telefunken Computer GmbH. After that it was IBM vs Siemens in Germany - both with basically compatible machines and many offerings to win over each other's customers :))

The very same happened during 1970s and 1980s with mini computers. Companies like

offered quite successful mid size systems of their own design. All vanished one or the other way during the 1980s.

Finally repeated with micros(*4) during the 1970s and 1980s with names like

  • Diehl,
  • PCS or
  • Triumph Adler (Alphatronic) (*5)

In the end even Siemens sold off its computer business to Fujitsu.

I'm sure similar history can be collected for other countries/companies/designs. Bottom line, seeing the demise of Soviet designs solely as losing in East/West frame may not tell the whole story.


P.S.: On purpose only 3 examples each - there were many more.


*1 - The first ever fully transistorized computer

*2 - Not only the fastest computer designed in Europe in 1969, but also the first operated with a mouse.

*3 - Their 600 series is may be one of the most interesting 8 bit designs I've ever studied. One may describe it as a 6502 on steroids with handy mixture of x86 and /360 like string instructions.

*4 - Independent micro processor development also happened in Germany - like AEG's CP3F, better known as it's Fairchild F8 offspring.

*5 - Which BTW bought Pertec who had swallowed MITS before.

2
  • Where can I find out more about the Dietz 600? Commented Aug 29, 2023 at 3:12
  • @Lorraine Not much on the net. Bitsavers ha the hardware documentation for the latest (ca. 1977 introduced) 621-X series CPU, and there are a few bits around the net but mostly in German and French. Manuals exist in German, French and English. I only own German ones.
    – Raffzahn
    Commented Aug 29, 2023 at 11:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .