A grain of truth in the crazy statement “there are no and cannot be chips in Russia” and what follows from it

There was a dispute a couple of days ago on Facebook with sociologist Alexey Roshchin, in which he made a completely insane statement “there are no and cannot be chips in Russia, and if there is something, it is two generations out of date”. It is clear that there are chips in Russia, for example, the MIK32 AMUR microcontroller, released in Zelenograd based on the processor core from the St. Petersburg company Syntacore. Moreover, if you compare its Amur with the STM32 U0 2024 (a low-power microcontroller from the world leader at 160 nm, 56 MHz), you can't say that the Russian one is “two generations out of date”. The coolness of microcontrollers is not in nanometers (a 3 nm chip will not survive next to a hot car engine anyway), but in system and microarchitectural solutions (tricks for saving dynamic power consumption, efficient DMA, even AI extensions in the style of ARM Ethos-U55).

Considering that Roshchin is not a crazy person, I tried to understand what he means, and I think I did. The world of microcontrollers and embedded processors does not exist for him (humanitarian sociologists do not think that microcontrollers with RTOS are now in irons), and chips for him are computers with Windows, which are used by “everyone”.

And if you look at it with such a definition, then I completely agree with Roshchin. In Russia there are not, have not been and never will be competitive processors of the x86-64 architecture on which it is possible to run Microsoft Windows normally (i.e. without slow simulation with QEMU and without binary translation). The unsuccessful benchmarks of Elbrus in Sberbank are a clear demonstration of the dead-end nature of such an idea.

But the main point here is not the lack of expertise in designing processors in Russia (although there too), but that the locomotive of x86-64 technology together with Windows has long since left. For Windows – 30 years ago, for x86 – 40 years ago. Although x86 was rejuvenated a little in the 1990s – due to the introduction of a dynamic pipeline in PentiumPro in 1996 and 64-bit from AMD later.

But trying to repeat all the dead-end solutions and undocumented compatibility that have accumulated in x86 since 1978 / Intel 8086/8088 (which in turn contains now meaningless compatibility with 8080 since 1974, for example H/L register) – even with perfectly trained engineers, this would mean thousands of engineers picking around in all sorts of nonsense for many years and still would not lead to success. The same kernel and Windows applications.

The path of Russia, China and other emerging global alternatives is Linux computers on RISC-V. There you can legally, constructively and effectively use open solutions from all over the world, on the basis of which you can build your own.

And x86-64 with Windows are dying technologies, despite billions of users in the world and despite the fact that they will die for another hundred years. Raising your eyes to the ceiling that Russia does not have this and will not have it is like scolding Nigeria that it cannot copy computers on the IBM Z processor. Haven't heard of this? This is a modern continuation of the IBM 360, which the USSR copied as the ES computer in the 1960s. Yes, it is still alive!!! It works in banks that have not left this locomotive for 80 years. But there is no need to repeat it.