The birth of the Soviet missile defense system. BESM-6. Outcomes

25
The birth of the Soviet missile defense system. BESM-6. Outcomes

Basic computers BESM-6. Source: JINR photo archive

Another myth


Another myth is that the unique developments of BESM-6 were ruined by the transition to the soulless EU.

In fact, as we have already found out, BESM-6, in principle, could not occupy the niche of a universal computer for the State Planning Commission, research institutes, factories, etc. Yes, it was conceived as such and in theory it could be, the goal of the path was absolutely correct, but the path turned out to be crooked and led out in the wrong direction. BESM-6, in principle, was not a competitor to the EU, as a mining dump truck is not a competitor to the universal cargo platform, especially since the dump truck came out of it so-so, the older EUs were not much weaker, but much more manageable.



The BESM-6 project was not closed at all with the transition to the EU, a total of 367 cars of different versions were produced, they were made from 1968 to 1981, in the early 1980s a version of the BESM-6 was produced on the IS - Elbrus-1K2, then the next version was - Elbrus-B. BESM-6 of all types were in operation until 1993-1994. Moreover, in 1971, work was initiated on the design of BESM-10, but they were curtailed for reasons beyond the control of the EU series (Lebedev's death, intrigues in the USSR Academy of Sciences, general collapse that began in the late 1970s, knocking out money by competing groups of developers under "Elbrus", etc.).

The EU did not prevent Yuditsky from developing his minicomputers, Kartsev from building the monstrous M-13, and so on. She did not interfere with Elbrus either. Moreover, a considerable number of original cars were also developed under the EU label, we will talk about them later. In addition, as we have already mentioned, BESM-6 failed to cope with the main task - launching all the software that was written for CDC machines.

It was the failure in this that showed that if you want to use stolen software without problems, you must completely copy the architecture of the machines. Emulation is inefficient and of limited applicability, and cross-translation is difficult and time consuming to debug.

And before ITMiVT and Lebedev personally in 1959, the task was set - to make it possible to run American and European software on Soviet machines. Ideally, no problems at all, in reality - with recompilation and not too time-consuming debugging, because writing everything of your own was a utopia, this became clear already by the beginning of the 1960s.

It's not even about operating systems and compilers - they somehow made their own, it's about application packages, of which there were many thousands already in the 60s, and no Soviet research institutes were able to provide software of such quality in such quantities. Only to copy, only in this way it was possible to provide the national economy with modern software and not to lag behind completely.

The last two myths


For a snack, we just have to disassemble the last two myths.

The first of them, in general, is not a myth; rather, it is story, which until now has not been fully told in Russian-language sources, and the author wants to correct this omission.

As you know, to the great regret of all technoarchaeologists, amateurs and researchers of the history of computers, the USSR (there is nothing to say about Russia) did not take very good care of old machines. A sad fate befell almost all of the domestic computers, and now we can only touch the crumbs of the entire technical heritage of those times.

MESM was melted down for scrap metal, only one console remained from the Setun-70, parts of the processor from the "Electronics SS BIS", a couple of spare parts from "Strela", some Elbrus-2 boards can be seen in California in the greatest computer museum in the world Mountain View Computer History Museum. The remains of the only CDC Cyber ​​170 in the USSR are at the SPII RAS in St. Petersburg, while the only one in the Burroughs Union has disappeared without a trace.

From more than 300 BESM-6, almost nothing survived, in total, the boards of each machine contained more than a kilogram of precious metals, so their fate in the late 1980s - early 1990s was a foregone conclusion.

There is one copy in the Polytechnic Museum in Moscow, but its completeness and performance are questionable.

Nevertheless, there is the only place in the world where a complete BESM-6 has been preserved in absolutely working order and it is located in Great Britain - the famous The Science Museum in London. The car was rescued from the collapsed Soviet Union in 1992 by old computer fanatic and explorer Doron Swade during a tumultuous time when it was possible to buy and take out even a tank, much less an ancient computer.

Technoarchaeologists of future times will be eternally grateful to the persistent curator of the Department of History of Information Technologies of the Museum of Science for this feat (by the way, he, in general, is the discoverer of Soviet computers in the West, and he also pushed the posthumous introduction of Soviet scientists, including Lebedev, into the IEEE Hall of Fame, presenting their work for the Babbage medal).

What is the myth here?

It is that from the entire long article about Sueid's journey to the wild Siberia of the 1990s behind the legendary computer, in Russian-language sources exactly one sentence is quoted and quoted:

Russian BESM-class supercomputers, developed over 40 years ago, refute US claims of technological superiority during the Cold War.

In fact, this story is much more interesting, therefore we will present to the readers other parts from his fascinating anabasis "Back in the USSR A museum curator suggests Russia's BESM supercomputer may have been superior to ours during the Cold War".

November 18, 1992 Wednesday. Deep in the heart of Siberia lives a winking monster. It is rumored to occupy an entire floor and is equipped with thousands of flashing console lights. It was once part of the military, space, engineering, meteorological and computer programs of the USSR and is the last working version of the legendary supercomputer BESM-6. He stands among the wreckage of three of his compatriots, which were shattered and melted down to recycle their precious metals.

I came to the former Soviet Union to save the last car from the same cruel fate.

We arrived in these snows and fierce winds to get a working supercomputer BESM-6 for the National Museum of Science and Industry in London. In preparation for the trip, I read everything I could find about Russian computers. The literature search was both puzzling and revealing.

I learned that Russian computer culture has its own icons: Ural, MESM, Ryad, Nairi, Strela, BESM, Elbrus - acronyms with the same rich history and personal associations for the Soviet computer community as our abbreviated mantras are for us. However, due to secrecy during the Cold War, these machines are virtually unknown to Western computer historians and hardly ever mentioned in the historical canon ...

I am interested to see the legendary machines that I read about - Ural, MIR, and especially Elbrus, a Burroughs-based supercomputer that replaced BESM. Shortly after our arrival, I turn to one of our hosts, Dmitry, a young computer scientist from the institute, who will be our main translator, and inquire about these historic computers. He answers my questions with an empty look and politely shies away, so I leave this topic.

We begin several days of intense negotiations on the price and delivery procedure for the historical equipment that we have come to buy - BESM, Kronos workstation, AGAT personal computer (Russian Apple II) and other machines. Consent on each item is accompanied by an obligatory sip of vodka. On the third day, our incessant meeting schedule suddenly changes. For no apparent reason, Dmitry announces: At 15:30 you will see Elbrus.

This is how I comprehend the main principle of doing business in Russian: what matters is not what you do, and not the level of your power; what matters is who you have established personal connections with. Three days of negotiations seem to have established the necessary confidence. Now our owners cannot refuse us.

November 21, 1992. Saturday. We need a break. We were hoarse and stiff from hours of conversation and accompanying booze. Dmitry and three of his friends from the institute take us with them to a huge flea market that works all year round in the frozen wastelands near Novosibirsk. The market is called barakholka, which literally means "garbage place".

We were told to hide our excursion from the directors of the institute: they are nervous about the risk to foreigners from hostile local residents. Dmitry warns us not to have any money or cameras with us and in no case speak English. If we want to buy something, we must signal and leave so that we will not be heard. Our companions from the institute will do business for us.

The temperature is well below zero and there is light snow. Next to livestock, car parts, furs, frozen meat and household goods, we see counters with integrated circuits, electronic components, peripherals, radio components, chassis parts and assemblies - Siberian Lyle Street in the open air.

Among the loot - converted clones of Sinclair ZX-Spectrum with Russian documentation and games stored on audio cassettes. Clones come in a variety of shapes, colors and designs and bear little resemblance to their Western counterparts. Their motherboards were informally made in government electronics factories by workers who then assembled computers at home and sold them one or two at a time, privately or at flea markets.

We end up buying two Sinclair clones; one of them comes with a guarantee - a handwritten note with the phone number of the teen who put together the device. Cost: equivalent to $ 19.

We return to the institute with our treasures. Once inside, I am amazed at the contradiction: the abundance of personal computers in the building contradicts the rules established by CoCom during the Cold War - the rules that restricted Eastern Bloc countries from acquiring advanced technology in the West. I mention this to Dmitry.

Yellow PCs, he laughs, waving his hand at the secretaries' color screens. He explains that these computers are unbranded machines, acquired through deals with factories in East Asia under contract with Western companies.
“So,” I say, “the Russians have the same passion for personal computers as we do?” In response, Dmitry points to the barred windows of the institute. "What do you think is the distance between the rods?" He asks.

I stare back in bewilderment.

“Slightly less than the width of a computer,” Dmitry replies. He assures me that he is serious and explains that the bars were installed to prevent computers from being stolen from the windows. But something still puzzles me. How, interestingly, does this combine with what I noticed outside the walls of the institute? There is a Russian abacus - schyotti - next to the cashier in most shops and hotels in the country. Sellers and clerks settle on it and then enter the total into the cashier, although most cash registers can add automatically.

When I ask Dmitry about this strange practice, he explains that the population does not trust new technologies, and schyotti is a symbol of a traditional, trusted procedure. Paradoxically, schyotti is now threatened by rampant inflation: traditional wooden frames and wire bridges cannot hold enough beads to cope with the smaller denominations of an increasingly depreciating currency.

November 23, 1992. Monday. The time has come to conclude our negotiations on BESM, arguably the most influential computer in the history of Soviet computing. These giant machines - from the prototype, BESM-1 (1953) to the latest model, BESM-6 (1966) - were the workhorses of scientific and military computing, and the institute's four BESM system at one time supported more than 300 independent users.

BESM-6 is of particular interest: according to some sources, this is the last domestic computer, which in terms of performance was not inferior to its Western counterpart - the Control Data supercomputer of the mid-1960s. More than 350 BESM-6 were built. The last ones were decommissioned in the early 1990s.

Our negotiations to buy a supercomputer were excruciating, but ultimately successful. The system we deliver home includes a complete BESM processor, power cabinets, many peripherals, cables, documentation, and spare parts.

With a more detailed view of this outstanding Soviet supercomputer, we may be able to revisit Cold War claims about Russia's alleged technological backwardness and dispel or confirm some myths about the technological prowess of our new contingent allies.

As you can see, Sueid's quote, to put it mildly, is taken out of context, for all his love for computers, he never and nowhere claimed that BESM-6 surpasses everything that was created in the West. He just assumed that the study of this machine would be able to answer the question - whether America lied about this superiority during the Cold War.

Unfortunately, we do not know what answer he received when he brought home a precious car and examined it, but I think the answer is already obvious to the readers of the article.


Professor Tomilin at the Science Museum in London next to BESM-6 rescued from Siberia, photo from Tomilin's archive

We left the very last myth for a snack.

It is so popular that it is found everywhere, even in the Russian-language Wikipedia.

The computing complex, which included BESM-6, in 1975, during the Soyuz-Apollo space flight, processed telemetry in 1 minute, while the American side spent 30 minutes on such a calculation.

Its primary source is the only interview of an elderly BESM-6 programmer, Professor Tomilin (one of the authors of the same D-68 protooperative system), who sadly died quite recently in 2021.

Recalling his youth and work in the MCC in an interview with indicator.ru, he said:

I was located directly at the terminal of the complex, where the results of the analysis of the quality of measurements were reflected. There were the most beautiful measurements! From this terminal, from the peripheral machine of the AS-6 complex, I transmitted information about the quality of measurements to another floor at the BESM-6 via a loudspeaker.

The received information about the quality of measurements was followed from there: "Yes, let's take it!" , and the instructions of the operators were immediately transmitted to the billing programs).

As a result, the calculations were performed 20 minutes faster than those of the Americans (the results coincided), to which it followed from Houston: “How so ?! What kind of cars do you have? "

The solution was obtained faster due to human-machine interaction.

In general, according to the story of an elderly veteran, it is difficult to understand what, in principle, was happening there, so let's try to dig the situation from the other side and look at the NASA MCC to find out which mission control computers they used.

Thanks to the Soviet Union


In fact, the most amusing thing is that the Americans have to thank the Soviet Union for the development of astronautics. It was the launch of Sputnik-1 (which no one expected from the USSR) that led the United States to shock for some time when it saw a clear gap in its technologies. After such a savory kick in self-esteem, three months later, the famous agency for advanced defense research DARPA was created (in its modern form), and six months later - in the summer of 1958, and NASA.

At the same time, for some time NASA did not have a colossal budget and some kind of extreme technologies, until 1958, the Jet Propulsion Laboratory (JPL), which was responsible for early experiments with rockets, generally managed with the staff of "human computers" - girls-computers armed with comptometers , tabulators and later - the old IBM 1620.

The use of human computing stations, in general, was widespread in the United States in certain areas no less than in the USSR, and this practice ceased only after the infusion of colossal funding in the wake of the "Sputnik" flight.

Let's open the book Computers in Spaceflight: The NASA Experience and see what the complex of several BESM-6 competed with:

America's most impressive contribution to the International Geophysical Year (1957–1958) was Earth's Vanguard satellite. In June 1957, the Vanguard Project on Pennsylvania Avenue, Washington, DC, established a Real Time Computing Center (RTC) consisting of an IBM 7044 computer. A 40 instruction computer program developed for Vanguard was used to determine the orbit in real time.

Thus, IBM gained early practice in the basic skills required for flight control. In 1959, when NASA was about to sign the contract to build the MCC for Project Mercury, IBM had experience to refer to in its proposal, as well as a working computer system from Project Vanguard.

On July 30, 1956, NASA entered into a contract with Western Electric to develop tracking and ground systems to be used at Mercury, and by the end of 1959, IBM was subcontracted to supply computers and software. The location of the computer system remained Washington.

The following year, NASA founded the Goddard Space Flight Center, and since it was less than half an hour from downtown Washington, placing computers there provided the same infrastructure benefits. The combined NASA and IBM teams used the old computer system in the city center until about November 1960, when the first of the new computers for the Mercury, the IBM 7090, was ready for use at Goddard.

NASA's James Stokes recalls that when he and Bill Tindall first walked into the new computer center, they had to cross a dirty parking lot to reach a "building" with plywood walls and a tarpaulin top, which was disconcerting IBM engineers trying to keep the system up and running in the field. This building became the third building of the new Space Flight Center.

The IBM 7090 central computer was the heart of the Mercury control network. In 1959, the Department of Defense challenged the computer industry by ordering a machine to process the data generated by the new Ballistic Missile Early Warning System (BMEWS).

IBM's answer was the 7090. Essentially an enhancement to the 700 series (used to develop the Mercury), the 7090 used a new I / O concept pioneered in the 709 and was so large that it needed up to three small IBM computers. 1410 for input and output control only. The Department of Defense's BMEWS needs matched Mercury's needs for data processing and tracking.

To provide the reliability required for manned flight, the basic configuration of the Mercury consisted of two 7090s operating in parallel, each receiving input, but only one capable of transmitting output. Called the Mission Operational Computer and the Dynamic Standby Computer, they migrated to the Apollo program and became NASA's first redundant computer system.

Switching from the main computer to the backup was carried out manually, so the decision was made by a person. During John Glenn's orbital flight, the main computer went out of order for 3 minutes, proving the need for an active reserve.

Three more computers were later added to the Mercury network. One of them was 709, designed to continuously predict the impact points of rockets launched from Cape Canaveral. He provided the data needed by the security officer at the range to decide whether to interrupt the mission.

Another 709 was at a tracking station in Bermuda with the same duties as a couple of cars in Goddard. In the event of a communication failure or double failure of the central computer, it became the main computer of the mission. Finally, a Burroughs-GE guidance computer radio-controlled the Atlas rocket during its ascent into orbit.

The deployment of computers outside Washington and the deployment of flight control personnel at Cape Canaveral created a communications problem that found a unique solution. In early digital computers, all input data entered memory through a central processing unit. Large amounts of data that needed to be received in a short amount of time often accumulated while waiting for the processor to handle the stream.

The solution to this problem was direct memory access over data pipes, pioneered by IBM in the 709 and later in the 7090. By using pipes, data processing could continue during I / O, increasing the overall system throughput.

The Mercury 7090 systems were four-channel. Typically, I / O peripherals are connected to channels that are physically close to the machine, but the peripherals (plotters and printers) controlled by Mercury computers were about 1 miles away in Florida. The solution was to replace channel F on the 000 with a dedicated IBM 7090 channel coprocessor.

Four subchannels shared the data processed by the 7281. One was Burroughs-GE's input for data used in powered flight path calculations. The second entered radar data to determine the trajectory and orbit. Two output subchannels controlled displays at the Mercury Operations Center at Cape Canaveral and locally at Goddard.

These points were connected by a land line, which made it possible to transfer data at a speed of 1 kb / s, which was phenomenal for its time. Distance and newness of equipment sometimes caused problems. From time to time during the countdown, data such as the one-bit take-off indicator was distorted and gave erroneous signals.

In most cases, such signals could be verified against other sources of information, such as radar data that contradicted the takeoff report. It was also common to see a delay of up to 2 seconds on the displays in the control center. During powered flight such delays could be significant; thus, there was a need for a separate forecasting computer and another machine in Bermuda.

In addition to flight control equipment, IBM has significantly advanced the theory of real-time operating systems by developing a complex of control programs called IBM Mercury Monitor.

To develop the control software package, IBM engineers had to work closely with NASA specialists who knew the subtle details of the mathematical definition of orbits, and they also brought in Professor Paul Herget, director of the Cincinnati observatory.

When the Mercury program was completed in 1962 and NASA began accelerated preparations for Gemini and Apollo flights, the agency decided to place the computers at a joint center in Houston. For IBM and NASA, the development of the Mercury control center was very beneficial, the IBM Mercury Monitor and Data Communications Channel were the first of their kind and laid the foundation for many computer technologies.

Future multitasking operating systems and preemptive control programs owe their introduction to the Mercury Monitor, terminal-access mainframes such as flight booking systems are based on long-distance communications between Washington DC and the Florida spaceport. For both organizations, the experience gained by in-house engineers and managers has directly contributed to the success of Gemini and Apollo.

Even before the first Mercury orbital flight, NASA's mission control engineers were trying to influence the design of the new center in Houston. Bill Tyndall, who had worked at NASA from the beginning on ground control, realized that the deployment of space task force leadership at Langley Research Center, computers and programmers at Goddard, and flight controllers at Cape Canaveral posed significant communication and efficiency issues.

In January 1962, he launched an information campaign to bring all the components together in one place, at the new Manned Spacecraft Center. In April, Philco Corporation's Western Development Laboratories began investigating the requirements for a new MCC, one of the requests being to ease the work of dispatchers by installing equipment to display trajectory graphical information.

As a result, Philco developed a new concept of flight control, describing literally everything from physical computers to information flows, displays, reliability research and even software development standards, indicating that modularity of programs is essential.

The final specification required a 336 hour mission without fail with a 99,95% probability. To achieve this reliability, Philco examined existing computer systems from IBM, UNIVAC and Control Data Corporation, as well as its own Philco computers 211 and 212, to determine what type of machines were needed and how many would be required.

As a result of calculations, three possible configurations were obtained: five IBM 7094 (direct successor to the 7090 with the best operating system IBSYS); nine UNIVAC 1107, IBM 7090 or Philco 211; four Philco 212; four CDC 3600.

Regardless of which solution was chosen, it was clear that the complexity of the Gemini-Apollo center would be much higher than that of its two-computer predecessor. To make the system as inexpensive and simple as possible, NASA indicated to potential bidders the need to use off-the-shelf equipment.

IBM quickly responded to NASA's proposal and in September unveiled a 2-inch-thick folder with hardware and software suggestions, including a detailed list of the staff they were going to bring to the project. While the company knew it was the lead candidate (Tyndall's endorsement could hardly have gone unnoticed), it carefully negotiated specifications, such as making it clear that unit testing would be the norm in software development.

However, there was one area in which their document differed from Philco's calculations - the number of cars required. Perhaps to lower the total price, IBM offered a group of three 7094 computers. They suggested that if one machine was to run the orbit calculation program, the second became the control, and the third was the backup, then they would provide 97,12% reliability, and on critical sites to the desired 99,95%.

Eighteen companies have entered the RTCC tender, including powerful competitors such as RCA, Lockheed, North American Aviation, Computer Sciences Corporation, Hughes, TRW, and ITT.

As a result, NASA inclined, as you might guess, to the IBM proposal, they signed a contract until 1966 for $ 46 million (about half a billion in modern prices).

NASA's requirements for Gemini flight control software resulted in one of the largest and most complex computer programs in history. In addition to all of Mercury's needs, Gemini's proposed rendezvous and orbit changes have caused an almost exponential increase in the complexity of orbit determination software. Placing a computer on board the spacecraft has led to the need to use parallel computing as a backup, as well as the need to develop a way to use a ground-based computer system to update Gemini data.

IBM has responded to the increased complexity in several ways. In addition to increasing its headcount, the company has introduced stringent software development standards. These standards were so successful that IBM adopted them throughout the company during the development of key commercial mainframe software systems in the 1970s.

In more complex areas, IBM turned to expert consultants and sponsored a team of 10 scientists looking for solutions to problems in orbital mechanics.

Even with better tools and a more powerful computer, the computing power requirements quickly exceeded the 7094's capabilities. IBM recognized that a machine's 32K RAM would not be enough, so it suggested using pre-buffering.

Because of the size and speed requirements of the Gemini software, the commercial practice of using tapes for pending programs became impossible, so IBM upgraded the 7094 to the 7094-IIs with 65KB of main memory and an additional 524KB of additional ferrite RAM called Large core storage (LCS).

In addition, Philco's calculations turned out to be prophetic - even so the computing power was sorely lacking, and IBM increased the purely final machines to 5, as was originally predicted in the Philco specifications.

As a result, the programs from the tapes were pumped into LCS, and from there into RAM, the work on their docking laid the foundations for virtual memory technology - the main software achievement of the fourth generation of S / 370 machines in the early 1970s.

As the Gemini program continued, NASA became increasingly concerned about the ability of the 7094 computers to adequately support the Apollo program given the expected greater complexity of navigation problems. The real-time OS clearly needed significant improvements.

The demonstration of the project to President Lyndon Johnson turned into a shame, he arrived at the MCC, and NASA employees offered him to launch one of the flight programs. By chance, Johnson poked at the program, which had already been pushed out of RAM to tape, in the end, as those present described, minutes seemed to them hours while the president patiently waited for the download.

NASA decided to spit on IBM and buy the great CDC 6600 from Cray, whose monstrous computing power tenfold exceeded everything that was already installed in the MCC. The IBM deal was hanging by a thread, and as usual, they made a clever marketing ploy by promising to replace all 7094s with the much more powerful newest S / 360 mainframes.

The piquancy of the situation was that there was still six months left before the delivery of the S / 360, the car was not ready, but there was not a word about this in the press release. NASA sighed and withdrew the order for the CDC 6600. Cray sued IBM, claiming they cheated by claiming that the machine was unavailable at the time as finished, in order to oust the CDC from the market. There was nothing to cover, and IBM was fined $ 100 million for unfair competition.

As a result, for unmanned Apollo flights, IBM managed to replace only one machine, 4 remaining 7094 still continued to control the mission. It wasn't until 1966 that IBM finished developing a new real-time operating system for the S / 360 - RTOS / 360.

As a result, the Apollo manned flight was supported by two S / 360 machines, one working and one backup. This scheme lasted until 1974, when annoying IBM again won a tender for the supply of equipment for NASA from Computer Sciences Corporation. From 1984 to the mid-1980s, flight control, including the Space Shuttle program, was carried out by five System 370/168 mainframes. In the late 1980s, they were replaced by the IBM 3083 mainframes, which became the fourth generation of Mission Control machines.

During this time, the importance of ground vehicles dropped significantly, as spacecraft computers became fast enough and advanced enough to perform most of the trajectory calculations directly on board during flight. All these computers were also built by IBM: ASC-15 for Saturn 1, ASC-15B for Titan Family, GDC for Gemini, LVDC for Saturn 1B / 5, System / 4 Pi-EP for MOL and System / 4 Pi-TC 1 for Apollo Telescope Mount and Skylab.

Mainframe battle


So, in 1975, 2 mainframes IBM System / 360 model 95 met in the battle (special order from NASA, only two machines were created, an upgraded version of the model 91 with RAM on thin magnetic films, a more advanced and faster version of conventional ferrite memory, developed by Sperry for UNIVAC 1107 in 1962) from NASA and AS-6 in the Soviet MCC.


IBM System / 360 model 95 in all its glory at NASA. Photo https://ru.wikipedia.org

It should be noted that only one IBM machine was responsible for telemetry, and in fact the Model 95 was a true masterpiece.

It was announced as a direct competitor to the CDC 6600, the first IBM superscalar machine with full support for speculative execution, advanced cache, modern virtual memory, one of the first machines with multichannel RAM, the central processor consisted of five autonomous blocks: instruction block, real arithmetic block, block integer arithmetic and two channel coprocessors: one for RAM (actually modern DMA technology), and the other for I / O channels.

The advanced pipeline used IBM know-how - the dynamic instruction scheduling algorithm Tomasulo developed by computer scientist Robert Marco Tomasulo specifically for the S / 360. The algorithm can work with any pipeline architecture, so the software requires few machine-specific modifications. All modern processors, including the Intel Core line, use some form of modification of this method.

In theory, the model 95 overclocked to 16,6 MIPS (albeit with simple instructions), but this was already a stunning achievement by the standards of 1968 and remained so for general-purpose computers for many years. Comparable performance on microprocessors could only be squeezed out of the Intel 80486SX-20 MHz or AMD 80386DX-40 MHz of the late 1980s.

Honestly, in this battle, the unfortunate BESM-6 can only be pitied, but not everything is so bad!

As we have already said, with the general wretchedness of the element base and rather strange ones. mainstream development of computers, technical solutions, BESM-6 possessed a completely successful system architecture, allowing wide range of combining its computing elements, for this, the interface equipment was developed - AS-6.

The AC-6 was designed in a very tricky way. For its functioning, the available BESM-6 had to actually be disassembled into modules, and then assembled again as part of the complex through special switches.

At the first switching level, processors from BESM-6 and their RAM were connected using a specialized switching processor AC-6, obtaining what can now be called a symmetric multiprocessor architecture - up to 16 CPUs from BESM-6 with shared RAM. At the same time, during the assembly process, the processor cabinets were moved and reconnected to achieve minimum signal delays.


Actually AS-6 as it is, photo http://www.besm-6.su

The second level of switching included the channel coprocessors PM-6, which were so missing in the original BESM-6, connected into a network, through which various peripherals were connected.

Finally, the third level consisted of interface devices with external data sources.

All this was collected on the basis of channels from the EU mainframe (even the haters of the Unified System cannot but admit that he helped the old woman BESM-6 a lot). All additional AS-6 coprocessors were assembled on the same DTL as BESM-6.

The software had an extremely exotic architecture - its own operating system (OS of the same name AS-6) was responsible for CPU management, its own (!) Separate operating system (OS PM-6) was responsible for peripheral processors. If it seemed to someone that the scheme lacked insanity, we hasten to console you - individual BESM-6s in the complex worked under the control of their native OS to choose from (DISPAK, etc.).

The original was the AC-6 control processor itself, which is a deeply modernized BESM-6 (yes, BESM-6, which drove other BESM-6s). It was more powerful than the original, with a performance of up to 1,5 MIPS with 256 kilowords of RAM and, of course, could use, as its own, the RAM of all other BESM complex through a channel of 86 buses with a total transfer rate of 8 Kb / s. Naturally, all this canal economy had its own food - the so-called. block UKUP (device for monitoring and control of the power supply system). The periphery was also taken from the EU (where else to take it).

As a result, the MCC AS-6, in a certain sense of the word, emulated the System / 360 model 95 architecture, only assembled from separate blocks, and with processors of a very different architecture.

The capabilities of this monster rested purely on physical limitations - in practice, the AC-6 was never used with more than two controlled BESM-6s at once for an elementary reason.

Even such a configuration required an extremely huge turbine hall of 200 square meters (not counting the separately removed periphery) and a power supply of no less than 150 kilowatts. The final speed of this complex is not just difficult to estimate, but generally impossible, since, as far as the author knows, no one has ever launched direct performance tests on the AC-6 in a complete assembly.

The actual performance of each of the BESM-6 in its composition was about 0,8 MIPS, the AC-6 processor itself added 1,5 more, it was unrealistic to compare this with the S / 360, since the architectural machines differed in every way - from the machine word (50 bits versus 36) before arithmetic (three parallel purely real processors were compared with separate real and integer).

In principle, if we take into account really high-quality mathematical and software and accept that in 1975 only one S / 360 counted telemetry versus three BESM-6 operating in parallel and the data was previously processed by a bunch of PM-6 coprocessors, it can be reasonably assumed that the speed of the AS -6 in the end was not inferior to the IBM machine and (with a certain stretch) could even surpass it.

We are not sure that the difference was exactly 20 minutes (and this does not mean that the everywhere-found version of the BESM-6 bike worked 30 times faster than the best American computers), but, perhaps, such a configuration could really compete with the CDC 6600.

Here are the memoirs of one of the employees of the Soviet MCC about those times:

In 1975, in the MCC-e, flight control of spacecraft of the Soyuz and Salyut type and flight under the Soyuz-Apollo program was provided by the AS-6 computer complex, consisting of 2 computers BESM-6 and 4 or 6 peripheral machines PM-6 (I don’t remember how many, I don’t want to lie, if my colleagues are still alive, please specify). All processing was carried out in real time. The PM-6 machines were connected to the transmission lines of telemetric and ballistic information and carried out its primary processing, at the rate of receipt. At BESM-6, the main processing of information was carried out, the formatting of working personnel and its issuance to monitors in the control room, to the ATsPU - upon request, and of course it was recorded on external magnetic media.

In the dry residue


The bottom line is this.

BESM-6 turned out to be slow by the standards of 1970, at the level of computers in 1959-1963. It turned out to be expensive and low-tech, hand-assembled from hundreds of thousands of discrete elements.

It was very specific in control and was only suitable as a number crusher; it was extremely inconvenient and hardly possible to use it as a universal or control computer. It was enormous in size and consumed a huge amount of electricity, again due to the element base that was outdated by 10 years.

And finally, it was completely unsuitable for what it was assembled for: to become an analogue of the CDC 1604, a computer that can be replicated in thousands for all research institutes and universities and use the entire array of American code without suffering from cross-compilation and rewriting of everything.

That is why BESM-6 was released, albeit with a record, but insufficient circulation, they simply overstrained themselves to tinker with an expensive, slow and outdated machine, without the proper amount of software, although more modern EUs on a more perfect element base were riveted by the Soviet industry in thousands without the slightest difficulty.

Actually, the EU project itself began indirectly due to the fact that the idea of ​​BESM-6 did not take off in the form in which it was urgently needed. This is where Elbrus's legs grow - the BESM-6 was not suitable for the role of a real supercomputer, there was only one CDC 6500 in the country, and a car five times more powerful than the BESM-6 was demanded by many, from rocket engineers to chemists.

Was BESM-6 a bad machine out of context?

No.

In 1959, it would have become a great machine (if it had been created independently, of course), in 1962-1963 - an excellent machine for narrow tasks, in 1965 - a normal machine. In 1968 it was to be taken out of production and put into a museum.

With such a life cycle, BESM-6 would definitely enter the pantheon of the greatest computers in history.

This was prevented by two small details - firstly, in the BESM-6 bottling of 1967 there was too little original (and the unoriginal was too fancifully mixed), and secondly, it appeared that year, when it would be wise to finish producing it. , being 10 years late.

As a result, she was born dead, and only the heroic efforts of thousands of man-hours of her selfless users were able to breathe life into this strange cadaver.

Why did she become a legend?

Well, for starters - in principle, its architecture for scientific applications was not so bad, and if you subtract ten years - it is not bad at all, even though it was not really good for anything else.

If we take into account the Soviet lag (growing every year) in the field of computers (in 1967 we were approximately at the level of 1959-1960 according to pessimistic estimates, 1961-1962 - according to the most optimistic ones) - BESM-6 was a masterpiece against the background of all kinds of " Nairi ", lamp" Ural "and other zoo of slow wretched designs originally from 1950.

In addition, it was objectively the fastest computer in the USSR (apart from secret military projects, the same M-10 cut it like a tortoise god, calculations of plasma hydrodynamics, which took hours on BESM-6, were considered in minutes) and, most importantly, widely available : almost 400 installations are no joke! At the same time, in the most powerful version - two machines in parallel, working with 6 coprocessors through the AC-6, as we said, it could even compete with the S / 360 model 95, and this was serious.

It also played a role that scientific centers, which had previously seen only the darkness and horror of Soviet informatization, finally got their own powerful machine.

Henrietta Nikolaevna Tentyukova, head of the LCTA OMOED sector, recalls again (JINR weekly "Dubna" No. 34 (4325) of 11 August 2016, "When the cars were big"):

And we thought: they give us formulas, and we count. You type multiple-digit numbers, the machine is cracking ... Yes, the method of least squares in its purest form. In general, work is a quiet horror. Two months later, we rebelled: why did we graduate from the university? Give us at least some scientific advisor! We are told: look ...

At about the same time, Venedikt Petrovich said: go to Moscow, there is an electronic calculating machine BESM. And at the university we only heard about electronic machines.

The first impression, of course, is grandiose: the hall is huge, there is no literature. Input only from the remote control. I had an acquaintance there who knew how she worked, and he taught me. On it I tried my first program, with tracks. She then, by the way, was still working on mercury delay lines ...

And then Dzhelepov said: let's buy a car too. And we bought "Ural". 100 operations per second, the memory is all on a drum ... But what is a hundred operations per second for our Institute?

… We worked with punched tape. She was then, of course, not the same as now. For some reason, it dried up all the time and made a very strong rustle when rewinding from reel to reel. This is how you sit in the car at night, you are kemar (the engineer is sleeping in the next room), and suddenly you hear: rustling, my dear! Lord, if only it did not break! And for some reason, the data made their way on a film strip. And all the time we were afraid that it would catch fire.

But most importantly, there was no software. I need, for example, a sine - I write it in codes. Once again I need a sine - I am writing again ... I remember the first thing that Blabberyap said when he came: Lord, how do you work here? Well, we are working ... Well, let's make at least some elementary system!

We were already part of BLTP, JINR was formed. Bogolyubov was the director of BLTP. He loved to walk surrounded by his students: Shirkov, Logunov, Polivanov, Medvedev ... and we ran to look at our bosses. Once Logunov, he kept in touch with our calculation group, gave me the task to recalculate the results from a Western magazine on a computer. The task turned out to be interesting. This is how we started our big tasks. The next one, for example, took 400 hours of machine time at the Ural.

We also went to Moscow in the "Strela" to count, this car was still more powerful than the "Ural". For a long time, everyone rode in chorus, led by Govorun and Igor Silin. Each with its own deck. Time was given to us on the night from Sunday to Monday. We left on Sunday afternoon, worked at night, returned on Monday, slept well the next day, from morning to work. This continued until we rebelled.

The talker was terribly surprised: how? what? do you need time off? It never crossed his mind ...
By the way, there was no issue on paper on Strela. On the "Ural" at least a printout of numbers could be obtained, but on "Strela" here is a deck of punched cards for you, without any overprinting, and study the punched cards into the light! They had an autonomous device for printing there, but on Sunday it did not work, and on Monday morning we were already leaving. And all this continued until we bought the M-20.

And, no, my brothers, there was also "Kiev"! "Kiev" is an epic! It was something so awful. It never really worked. Lida Nefedyeva and I were seated for him, to write elementary functions for him; memory was very limited, we had to save every cell, and so Lida and I were refined. But "Kiev" did not work, although the people of Kiev were constantly remaking it and asked us to wait another ten or fifteen minutes, and we sat on it all Sundays.

Well, we bought an M-20. Already the sixties have gone, Algol appeared, Lida Nefedieva read us the first lectures on Algol. Yes, civilization has begun. Life has become easier. Time on M-20 was allocated by the timekeeper. She distributed it like this: here's your time from 12:02 to 12:04, two minutes. And to keep track of the time, there was an alarm clock on the car. And we twisted it back all the time. You come to the car, for example, at two, and there it is half past one, or even one o'clock.

The tape recorders were then non-interchangeable, on which one you wrote down, read on that one, and it is good if you can read it. The magnetic tape was "running away" all the time and was wound up all over the partition. In such cases, it was necessary to stand on a bench, hook the tape and quickly wind it back on the reel. While you are running, you take it out and wind it up - your two minutes have already passed.

And we were not allowed to the punchers. We were afraid that we would break. It was only later that Nikolai Nikolayevich insisted when he returned from CERN. So if something urgently needed to be corrected, you fly into the punching room, stick your head through the window and beg: girls, for God's sake! I have a car. And the compassionate girls rushed to the puncher.

These were the 1960s, JINR, the main and most powerful computer center of the country, working on problems of nuclear physics of world importance. Obviously, when BESM-6 appeared there - they were ready to literally kiss the car on every board - there was nothing better, it was a complete nightmare.

Another reason for the warm attitude to BESM-6 was that it was one of its own, dear (well, more precisely, none of those who worked on it were able to identify its prototypes, and even now few think about it), and the ES computers were clones, which beat by pride.

In addition, the EU was extremely, extremely difficult for the Soviet culture of production, even taking into account the fact that the USSR mastered the S / 360 only by the mid-1970s.

As a result, the first series of the EU worked simply horribly, and many of the following ones, too, given the fact that they were riveted in thousands in factories with fundamentally different cultures. If the EU was lucky and got the production of the eastern bloc - the GDR, for example, it was happiness. The culture of assembly of our southern republics was legendary, more terrible than the stories of Lovecraft.

In 2000 E. M. Proydakov met with Emmanuil Grigorievich Kneller, now the president of Istrasoft, and recorded his memories of the appearance of the Istra-4816 personal computer on a dictaphone. EG Kneller led a small group that developed this machine at the VNIIEM branch in Istra.

He recalled:

It must be said that Iosifyan wanted the production of the car to be mastered in Yerevan. We even went and negotiated with the factory that produced the Nairi computer. However, the technological discipline there was even lower than at the "Schetmash". When they took me around the plant, showing the production, the chief engineer told me: “Here they are - smart guys from the mountains. You ask them: "Why are you leading the wire like this, and not as it is drawn on the diagram." He replies: "What am I doing worse?"

Considering that people used to assemble unpretentious PCs in this way - imagine what monstrous shoals they allowed in the assembly of mainframes.

There were frequent situations when it was not possible to launch an EU-placed university for months or even years. BESM-6 was many times simpler, if desired, it could be repaired with a hammer, a soldering iron and a famous mother.

Charisma


And, finally, one cannot fail to note the last important component of the popularity of BESM-6.

Despite its outward severity, like its creator Lebedev, the car had a certain charisma.

A cozy semicircle of racks, rows of fervently blinking light bulbs, the relaxed-intellectual atmosphere of Soviet research institutes of the 1960s – 1970s - all this was close and pleasant to a whole generation of developers and users. Again, working at BESM-6 meant frequent business trips in search of software (including to Germany and Hungary), receiving guests (including foreign ones) and other entertainments of the intellectual elite. Of course, this is why many have the warmest memories of BESM-6.

So, it remains for us to tell about the epic with missile defense computers of Lebedev's second favorite student, Burtsev, but first we need to refute another popular myth, launched in the most pompous and artistic presentation by the Rodina magazine and Elena Litvinova in the article “Sergei Lebedev. Battle for the supercomputer. While they were calling from the stands to overtake and overtake America, he did it quietly and, alas, unnoticed by his country. " The most epic paragraph in it is this:

Perhaps the most difficult thing in the life of Sergei Alekseevich. Discussions about the further development of computing technology became more and more heated. Lebedev was sure that we had to go our own way, to create our own line of medium-power computers and a new generation supercomputer. Opponents proposed to create a number of compatible computers, repeating the American IBM system.
Lebedev objected harshly: "We will make a car out of the ordinary."
Out of the American ranks!
In the winter of 1972, Sergei Alekseevich lay with pneumonia when he learned that the decision to copy the American car had been made finally. He got out of bed and went to the minister to convince him not to make a mistake that would set the country back years. Lebedev waited in the waiting room for over an hour. The minister did not accept him.
Who benefited from this turn to the West?
Perhaps this story brought the death of Sergei Alekseevich closer. He was getting sick more and more often. Alisa Grigorievna and the children were on duty around the clock in the hospital. The outstanding scientist died on July 3, 1974.

In fact, everything was, to put it mildly, not so, and we will consider this issue further.


Lebedev solders a supercomputer. Drawing from Supercomputers magazine # 1, 2010. Tragicomic picture, reflecting the whole essence of the development of Soviet computers

Продолжение следует ...
Our news channels

Subscribe and stay up to date with the latest news and the most important events of the day.

25 comments
Information
Dear reader, to leave comments on the publication, you must sign in.
  1. +4
    20 December 2021 19: 19
    Interesting and informative! Respect to the author hi
    1. +5
      20 December 2021 20: 49
      I join thanks to the Author! The cycle of articles turned out to be wonderful !!!
  2. +6
    20 December 2021 19: 40
    As always, very interesting, many thanks.

    > From 1984 to mid-1980s, flight control

    Here, 1974, I see)
  3. +1
    20 December 2021 20: 29
    the older EUs were not much weaker, but much more manageable.
    As a user of BESM-6 and EC 1055, I can only say that this is not true. This EU sucked. BESM-6 in the 70s was quite at the level of solving gas dynamics problems, by the 80s, of course, it was already outdated. But again, the EU is crap.
    1. ANB
      +3
      21 December 2021 09: 11
      ... This EU sucked

      From the article it became clear why we are arguing about the quality of the EU. In our school there were Gdrovskys. They worked great.
      1. 0
        21 December 2021 18: 32
        In our school there were Gdrovskys.
        The Germans, I remember, supplied the Robotron dot matrix printers. Worked well, only swayed when typing. The quality depends on the tasks to be solved, if every 20 minutes of the counting there is a failure, then it quickly starts to get boring.
        1. ANB
          +3
          21 December 2021 23: 58
          ... dot matrix printers "Robotron".

          We had one. Compared to the ADCU, it typed slowly but beautifully. And the staff member was Robotron. At the secretary.
          There were no failures. We worked for days. 1033,1045. Then they put 1066, generally a fairy tale. Probably, you got the Armenian assembly. And there were Minsk PCs in 1840.
          1. +1
            22 December 2021 08: 08
            Probably, you got the Armenian assembly.

            May be. As the French "Iris" appeared and later the VAX, then all the programs were dragged there. Both machines were allegedly bought for the USSR Academy of Sciences.
  4. +1
    20 December 2021 20: 46
    RAM of all other BESM complex through a channel of 86 buses with a total transfer rate of 8 Kb / s.
    Is there no error with the dimension? 100 Bits per channel won't be enough?
  5. -5
    20 December 2021 22: 50
    Another rewriting of history, when the writer, in addition to facts, adds myths invented by all sorts of liberals of that time. What is the lampoon about the size of the cells of the window grill, if instead of the system unit, you could take out a bunch of expensive boards and parts and take them out in your pocket.
    1. +3
      21 December 2021 11: 21
      One does not cancel the other, all the more it is palpable and takes a long time to disassemble and unsolder the chips when they can be carried away at once. So it’s never a story about bars on the windows.
  6. -2
    21 December 2021 01: 12
    Thinned slander in especially large proportions. How can the software be "stolen" if it was not licensed at that time, besides, there was no license agreement with the West ??
    1. +2
      21 December 2021 07: 52
      It was quite licensed for itself, people were paid money for it (at least a salary). It was not accepted for us to pay for software.
  7. +1
    21 December 2021 13: 53
    The software had an extremely exotic architecture - its own operating system (OS of the same name AS-6) was responsible for CPU management, its own (!) Separate operating system (OS PM-6) was responsible for peripheral processors. If it seemed to someone that the scheme lacked insanity, we hasten to console you - individual BESM-6s in the complex worked under the control of their native OS to choose from (DISPAK, etc.).



    By the way, such a "zoo of monsters" can still be observed alive in the modern world, just no longer in the space field. I mean newfangled mining and stock computing.
    There, too, there is a fierce incessant race for money, no less intense than in space in the old days.
    And a similar situation gives rise to highly specialized software and hardware monsters-exotic, strange and surprising solutions. Equipment and technologies are becoming obsolete before our eyes. They will also need a specialized museum soon ...
  8. 0
    21 December 2021 17: 20
    Once upon a time, one damn lucky and quite intelligent American manager, to whom, ironically, a computer genius called him, said something like the following - no one needs operating systems, everyone needs application programs.
    His thought can now be continued - no one needs programs, operating systems and even computers, everyone needs the result of their work - services.

    The USSR simply could not attract as many programmers to the development of applied software as Western companies could. Yes, they did it by forcing thousands of programmers (algorithms, systems specialists, coders) to sweat until the seventh sweat. In essence, for a penny. After all, a huge amount of their time was wasted. But what was left became the standard. Fortran with math libraries, BASIC, RT-11, MASM, COBOL, Algol, ADA, IBM System OS 360/370, JCL, etc.
    The only thing left for the USSR was to turn all this economy into RAFOS, DUVZ, Primus, etc.
  9. 0
    22 December 2021 04: 15
    Commenting on the mention of the author of the article about the Nairi computer series. "Nairi -3" in the provincial
    The university was never able to launch after delivery from sunny Yerevan to the center of Russia under
    rain and snow by rail, on the "Nairi-2" I learned to work with
    the keyboard of the ancient letter - typing apparatus "Consul". And the language of algorithmic
    programming for these machines. "Fortran", "Algol" - at the seminars.
    Well, there still in the "computing center" and the Belarusian "Promin" were. Before writing off ...
    And in some departments, analog computers were also observed.
    Well, the world has moved on to a faster and more convenient digital phone.
    Until they began to buy or lick it, and about launches on the undisclosed Kazakhs
    polygon sites in Priozersk (Sary-Shagan station) and could not be heard from
    "stupid box".
  10. -1
    22 December 2021 15: 18
    The author is in his repertoire. In the material about BESM-6, more than half of the text has nothing to do with BESM-6 at all. In itself, the presentation of how things were at NASA is interesting, but not at all on the topic, since there was clearly no need to write so much to compare the affairs of "they" and "we".
  11. IVM
    0
    6 January 2022 14: 03
    I do not presume to evaluate the content of the text about the BESM-6 - there is practically nothing about this computer in the text. In addition to mentioning the 50-bit system (which is not true, BESM-6 was 48-bit). But there is an extremely disrespectful mention of A.N. Tomilin (an elderly BESM-6 programmer, Professor Tomilin (one of the authors of the same protooperative system D-68)). The "aged" programmer was one of the founders of IT in our country and it would not be worth mentioning him without the respect corresponding to this wonderful person. I did not find any substantial, analytical information about the architecture and capabilities of BESM-6 in the text. Very much like an anonymous libel ...
  12. 0
    7 January 2022 01: 20
    "Registers of registration" with offsets is, in general, segment addressing, plus or minus characteristic of this generation, but there is still a lot of things missing before real virtual memory.
  13. 0
    11 January 2022 22: 02
    It is necessary to resume the production of BESM-6! Most importantly, it was entirely on the domestic element base. And not from Chinese boards, like all sorts of "Lomonosovs" there.
  14. 0
    11 February 2022 21: 51
    The greatest respect to the author!
  15. 0
    11 February 2022 22: 06
    There is a suggestion to gather around the fire with tea (well, or ...), and remember those times of blue robes (in our KIVC, they were the dress code), when everything was at the level of the experiment ...

    https://invite.viber.com/?g2=AQBfxNvJItLKU066aO2xxM2Fkkmd2XLpVTIujrQIXmzAmDvgHFKQbTDGqPiFy1v%2B
  16. The comment was deleted.
  17. 0
    18 February 2022 16: 06
    In general, the author famously crap one's pants for more than half a century of work of our domestic programmers and electronics engineers ... Probably that's why they are so grateful to him
    Quote: Alien From
    Interesting and informative! Respect to the author hi

    Ba-a-a-l-shoy! Better late than never ... to understand to them that their life was lived in vain and they had to either leave on time or change their profession.
  18. 0
    27 February 2022 12: 50
    The author is 100% anti-Soviet! If only to crap everything Soviet and exalt everything American! To crap BESM-6 and Elbrus-2 - well, that’s all right ... Suffice it to say that Elbrus-2 is still successfully providing missile defense for the Moscow region. And to crap the great Lebedev... It's the same as to crap Kalashnikov, Korolev or Tupolev!
    1. +1
      1 March 2022 11: 50
      Quote: Alex968m
      The author is 100% anti-Soviet! Just to crap everything Soviet and exalt everything American

      I don’t know ... maybe they pay him just for this, but for a normal person, the serial chewing on what “was not like that” is not pleasant. And how it should have been, eh?
      Much worse is the strained, thoughtful show off of "specialists in electronics and programmers" declaring "respect" to the author, who poured stinking shit on their entire work activity in the USSR for decades.
      Author ! Give it some more ..... show that they have been complete nerds all their worthless lives .... they didn’t create anything worthwhile .... and if they received money, then let them get up .... where and how follows - lumpy.

"Right Sector" (banned in Russia), "Ukrainian Insurgent Army" (UPA) (banned in Russia), ISIS (banned in Russia), "Jabhat Fatah al-Sham" formerly "Jabhat al-Nusra" (banned in Russia) , Taliban (banned in Russia), Al-Qaeda (banned in Russia), Anti-Corruption Foundation (banned in Russia), Navalny Headquarters (banned in Russia), Facebook (banned in Russia), Instagram (banned in Russia), Meta (banned in Russia), Misanthropic Division (banned in Russia), Azov (banned in Russia), Muslim Brotherhood (banned in Russia), Aum Shinrikyo (banned in Russia), AUE (banned in Russia), UNA-UNSO (banned in Russia), Mejlis of the Crimean Tatar People (banned in Russia), Legion “Freedom of Russia” (armed formation, recognized as terrorist in the Russian Federation and banned)

“Non-profit organizations, unregistered public associations or individuals performing the functions of a foreign agent,” as well as media outlets performing the functions of a foreign agent: “Medusa”; "Voice of America"; "Realities"; "Present time"; "Radio Freedom"; Ponomarev; Savitskaya; Markelov; Kamalyagin; Apakhonchich; Makarevich; Dud; Gordon; Zhdanov; Medvedev; Fedorov; "Owl"; "Alliance of Doctors"; "RKK" "Levada Center"; "Memorial"; "Voice"; "Person and law"; "Rain"; "Mediazone"; "Deutsche Welle"; QMS "Caucasian Knot"; "Insider"; "New Newspaper"