Build FreeBSD on Linux for Alternative architecture.

Neither Linux OS (nor linux kernel build scripts these days) nor FreeBSD are suitable tools for Cross Compiling. If they were then doing certain cross compiles "would be an option" yet there is no option.

I'm not allowed to discuss it (Sir Dice? or who? would IP block me). But I've tried cross-compile linux - it's ability is destroyed for "typical hobbyists" (i'm not allowed to say how) and similar new management has put "CLANG" as the only compiler in FreeBSD which itself admits "it has no way to compile itself without the system it was compiled upon running it - and there again it can only compile the same version" (without massive difficulties) - which i've also tried. (more than one BIG TECH company is involved and at least one of makes cross compilers "that work with one click, and must be bought" - hopefully i've not said too much to swallow)

My advice is don't. By the time you finish BIG TECH will have made everything altered and irrelevant as it does in (let's say a planned manner).
Seriously gotta try that substance some day 🥴🥳

Ok, really seriously: Cross compiling is always the way to get something for a new hardware architecture running in the first place. Of course, both GCC and LLVM can cross-compile perfectly fine. I'm cross-compiling FreeBSD for i386 and armv7 on a FreeBSD/amd64 host using LLVM, and userspace applications for Windows and Linux on a FreeBSD host using GCC...
 
DJs3000, can you please contact me on arrowd@freebsd.org ?

SleepWalker, what you basically need is just 40k rubles to buy this: https://www.chipdip.ru/product/bfk3.1
Maybe we can arrange a crowdfunding, I'll hapilly donate.
About $521.52 USD or 459.71 euro... Right on par with some of the cheaper Threadripper boards.
--
eurohick2 : This project is most likely aimed at the Russian market, rather than the global one. TSMC is probably only too happy to have a customer or two for some of its under-utilized fab capacity.
 
About $521.52 USD or 459.71 euro... Right on par with some of the cheaper Threadripper boards.
--
eurohick2 : This project is most likely aimed at the Russian market, rather than the global one. TSMC is probably only too happy to have a customer or two for some of its under-utilized fab capacity.
Aparently, Baikal BE-T1000 is 28nm and IIRC Russia have a 28nm fab. [EDIT] But yeah, it seems they are working with TSMC.

Correction, based on early 2021 information the 28nm fab is being deployed yet.
 
Aparently, Baikal BE-T1000 is 28nm and IIRC Russia have a 28nm fab. [EDIT] But yeah, it seems they are working with TSMC.

Correction, based on early 2021 information the 28nm fab is being deployed yet.
at least they have a cool 404 page
1642712345661.png
 
There is a Baikal M laptop in the making:
(not sure if that is a render of nice aluminum/metal case or a crap plastic one)

and an Elbrus too:

Lets hope they will not f!ck with the industrial design part.
 
Kinda reminds me of Koukaku Kidoutai Arise: Alternative Architecture. I acknowledge that there's not that many parallels between the anime and the technical process of photolithography development beyond the names. But, being forced to basically re-invent chipmaking from ground up since 1980s has to be pretty frustrating even for the bright minds that are at top of the academic field domestically. Small wonder the current situation lags 15 years behind the rest of the planet. Here's a comment from that article:
Translation:
Hot new popular idea that products and technologies have to [exist and be made] in Russia, especially on par with the rest of the world, that's irritating to no end.

This is the same crowd that never learned to mask up after 3 years, sitting around the kitchen table and complaining how the country has shortchanged them. Hella unique people.

Don't bash me over the English translation. You have to know both languages to understand the translation correctly, and realize that snarky attitude in the comment is normal and is part of the translation. 😩
 
What, there are posts after 23.02.22 there? It's dead in the water, guys.

IDK what you are asking about exactly, but if this is about the progress of microelectronic industry there were several news about several correlated points/topics but nothing solid or a complete plan is public known, as far I'm aware. In short, it seems the whole plan is being re-organized and some parts of the production process are being given priority. Example.

Apparently, a few days ago Malaysia offered to join the effort somehow.

You can follow related news in a centralized way in here.
 
You found a dedicated English-language vatnik forum? Impressive. Not in a good way, mind you.

You see, Elbrus is not really intended to be used in the military or industrial hardware, it's designed as a server/workstation processor. It wasn't really competitive at all in that niche at the TSMC 28 nm process and it certainly won't fare better if MCST backports it to some domestic 65+ nm process. This would probably give us an equivalent of Raspberry Pi Zero performance. Might as well just buy those from China.
 
Last edited:
You found a dedicated English-language vatnik forum? Impressive. Not in a good way, mind you.

You see, Elbrus is not really intended to be used in the military or industrial hardware, it's designed as a server/workstation processor. It wasn't really competitive at all in that niche at the TSCM 28 nm process and it certainly won't fare better if MCST backports it to some domestic 65+ nm process. This would probably give us an equivalent of Raspberry Pi Zero performance. Might as well just buy those from China.

Even a Pi can run Mathematica. If you have the brains to use Mathematica, that is. :p
 
You found a dedicated English-language vatnik forum? Impressive. Not in a good way, mind you.

You see, Elbrus is not really intended to be used in the military or industrial hardware, it's designed as a server/workstation processor. It wasn't really competitive at all in that niche at the TSCM 28 nm process and it certainly won't fare better if MCST backports it to some domestic 65+ nm process. This would probably give us an equivalent of Raspberry Pi Zero performance. Might as well just buy those from China.
Elbrus processors are used in military hardware (including radars) and weapon systems since years[1], there is no need to use the latest processor in those appliances. Not ever 45nm is generally necessary but 90nm is more than enough. IIRC Elbrus-2SM is the most used one in the latest systems.
 
I'm not a fan of military applications, but even Uncle Sam relies on COBOL and IBM mainframes to run the COBOL programs. A lot of times, the reason for not upgrading is because the old fart in charge of EVERYTHING has a lot of distrust in "Dem newfangled things", and will not be swayed, even with concrete evidence that flies in the face of that behavior. And then other old farts are pulled out of retirement so that they can tell you how to run rusty 50-year-old equipment. Those consultants are probably half blind, and unable to remember what they had for breakfast, but they will be able to recall instruction manuals from 50 years ago. 😩
 
I'm not a fan of military applications, but even Uncle Sam relies on COBOL and IBM mainframes to run the COBOL programs. A lot of times, the reason for not upgrading is because the old fart in charge of EVERYTHING has a lot of distrust in "Dem newfangled things", and will not be swayed, even with concrete evidence that flies in the face of that behavior. And then other old farts are pulled out of retirement so that they can tell you how to run rusty 50-year-old equipment. Those consultants are probably half blind, and unable to remember what they had for breakfast, but they will be able to recall instruction manuals from 50 years ago. 😩

You don't need a lot of power to run Target Acquisition or EW systems, the ultra-heavy optimized algorithms are the real thing behind them.

Space technologies may benefit from smaller and lighter hardware but at the same time those are harder to shield, but ever 50 Kg saved make a difference in these applications.
 
I'm not a fan of military applications, but even Uncle Sam relies on COBOL and IBM mainframes to run the COBOL programs. A lot of times, the reason for not upgrading is because the old fart in charge of EVERYTHING has a lot of distrust in "Dem newfangled things", and will not be swayed, even with concrete evidence that flies in the face of that behavior. And then other old farts are pulled out of retirement so that they can tell you how to run rusty 50-year-old equipment. Those consultants are probably half blind, and unable to remember what they had for breakfast, but they will be able to recall instruction manuals from 50 years ago. 😩
Also, the process for getting new components (say for example, a new processor and platform) certified for use is both time-consuming and very expensive. So often it is cheaper to just go with the existing solution for as long as you can, even if it gets more expensive as time goes on.

Then you just swallow the economic "big bang" when the technical debt and scarcity of components catches up with you.

Not speaking for any government or military.
 
Btw, by the end of 2023 China will have finished a 100% Chinese 14nm fab. That means fully immune to sanctions, and in fact Russia could probably acquire one and have a fully functional 14nm fab ever before they finish deploying/developing their own 28nm fab - if they prefer to do that instead of just having their processor made in there.
 
I don't like how politics seems to find a way to stick its nose into economy and technology and hinder cooperation WAY on the other side of the damn planet. Ego and violence are making way too much noise. Politics demand transparency from economy and technology, but the other way doesn't seem to work. Pretty hypocritical, but this is the reality we're facing while trying to cooperate with each other. 😩
 
Factory for the laboratory research line of the 28 nm process technology in Zelenograd
 
Plans for the production of Russian processors are shifting to the right
 
Back
Top