Build FreeBSD on Linux for Alternative architecture.

If any of you want the idea of the original post to be possible, you may want to make
an attempt with a few suggestions:
1. Become involved in the qemu-developers mailing list.
2. Join the POWER foundation. You will have access to the assembly instructions.
3. Options one and two combined to create a virtual machine based on the Tyan
or other real life models.
4. Contacting Open SPARC to make virtual machines based upon Leon and
other processors with assisting and you assisting qemu developers.
5. Asking SBC manufactures if they would assist you in making
software translations of hardware products.

When Torvalds was working for Via, an i686 CPU was developed with load-store
functionality.
Fabrice Bellard took time to develop software emulation of CPUs known as qemu
based upon the bochs project.
NetBSD emulation on Minix now exists.

I deal with hypothetical and theoretical ideas in AI, green technologies, and cybernetics.
This topic is not my specialty; however, I am aware of studies from my interaction with
different communities - as are many of you here.

Did you really made assumptions and jumped to conclusions, don't you? :)

The subject have literally nothing to do with POWER and very little with SPARC[1]. This is about Elbrus architecture.

Anyway, there are people already working on it or at least they were.

[1] nothing in practice.
 
I have to agree with rigoletto@ 's assessment. It's unclear what qemu, POWER and SPARC have to do with development of a totally different architecture. I'm guessing that they are usable as reference designs that Elbrus can try and imitate.
 
A legal question for you:
Do you have the forensic evidence
to support your conclusions
about my cognitive processes?

Meh. First it was obviously a joke but now you just proved you assumed things and jumped to conclusions... Thank you.

My suggestion for Qemu is a well thought out idea.

Thread 78073/post-530921
In case anyone's interested, I got my hands on Elbrus QEmu emulator and managed to execute a simple assembly-written Hello World executable on qemu-bsd-user. This is not much, but at least some movement in this direction.

Do you have a physical device that has the Elbrus CPU on it?
Did any of you even bother to see what type of architecture it is?

Thread 78073/post-567262
Anyway, there are people already working on it or at least they were.

And yes, there is someone with hardware in hands too.

2. Join the POWER foundation. You will have access to the assembly instructions.

Can you explain in any sane way how, by any possible means, joining the POWER foundation would help accessing MCST's Elbrus instructions?

Again, you just made it clear. You obviously assumed things and jumped to conclusions, like everyone does every now and then.
 
I really admire your effort but as I said in my previous post, the people who is/was working on it already have access to Elbrus hardware.

One of them is arrowd who is a FreeBSD developer too. I suppose he would have contacted danfe@ at this point.

Anyway, you won't find it easily (but just because this is in Russian only) but, in short, MSCT cooperate (for free) with anyone willing to port things to Elbrus architecture, and IIRC the all necessary documentation is publicly available on their website, but again only in Russian.

At this point, in order to help, this is not just necessary to have the technical capabilities but this is also mandatory to have a good domain of the Russian language:

Thread 78073/post-489641

Write me an email djs500@mail.ru In Russian, I will tell you about the plans. When I can get the cross-compiler, we can start working. I think there will be a lot of work.

Thank you.
 
apparently you don't understand this architecture very well read this article in russian do the translation
 
I already had a look, in the development of this architecture,
 
This is a family forum, please. :D
I can expand shkhln 's abbreviation into a plenty colorful expression in Russian... But you gotta have a pretty good handle on the Russian language, AND its colorful expressions to be able to understand that turn in the conversation. And I can respond with a pun :p

Slovak!

But even with that, the Russian sourpuss attitude can get in the way of good communication and technical collaboration.
😩

I do disagree with 'Family' part of 'Family forum' assessment, though... The intended audience of this forum is for the technically inclined public... Yeah, being polite and professional is expected, and having commitments (other than the FreeBSD workstations) should be respected - but I don't see this as 'family oriented'. I think the forums do have a section for gaming and whatnot... Our section on security can be thought of as 'keeping the family safe', but it's a bit too heavy on technical details, and has very little mention of 'I gotta keep my baby's iphone behind the pf firewall' type of content you'd see on TenForums.com ;)
 
I suppose astyle was commenting about the colorfulness of the Russian language in itself, technically (as to speak), and not in relation to the human physical expression during its pronunciation or use. The own technicalities of the pronunciation of the Russian language hold possibilities of human physical expression, which basically make the language technical sounding or ever pessimistic sounding to some people.

If you have no idea of what I'm talking about pay attention to the Belarusian language which doesn't have this characteristic, and is warmer sounding and allows a lot of more human physical expression.

[EDIT]

Just to add one thing, Russian Language is a «stressed pressured language» or Russians have a «stressed pressured speech»[1], which means there is an automatically tendency to shorter the atonic part of the words in an automatic tentative to pronounce it using the same amount of time used to pronounce the stressed part, what make the technical sounding thing ever more pronounced.

A good example is Portuguese, this is «stressed pressured» in Portugal but not in Brazil, what in practice means the Portuguese eat a lot of vowels when speaking the words, but in Russian the vowels often isn't ever there (as to speak).

Eg. colesterol (cholesterol)

Brazillian pronunciation: colesterol
Portuguese pronunciation: colstrol or ever clstrol

[1] these terms may not be 100% correct with the term used by linguists, but in the worse scenario this is not far.
 
Just how off-topic did we get here? I can as well nitpick on proper usage of English, the forums' official language:
You may judge me so long as you agree to be judged with me.
Poor English. The proper term would be by.

Completely different meanings, which I leave as an exercise for TBM to figure out. 😩
 

Interesting point:
"According to Zhbankov, in order to meet most of the requirements, you just need to repackage the server without touching the processor and OS itself"
 

Interesting point:
I'm sorry, but you've got a pretty bad translation. The article merely says that the Elbrus processors just lack the power of Xeons, which is what it takes to handle the mission-critical applications that Sberbank needs to run, such as containers, VMs, Java, and the like.

Edit: A more careful reading showed that the Elbrus processors were actually declared to "Have potential", because the performance discrepancy is only by a factor of 2-3, not 20-30 as initially expected.

Edit 2: Found the point, the translation is actually accurate... Yeah, the software could use re-organizing, but the processor did show potential. The reason 'woeful inadequacy and inability to meet hard requirements' is even mentioned is to set expectations for development of the processor itself.
 
The major problem was the server packaging not the processor. They tested the Elbrus 8S, need to wait the test with Elbrus 16S and an adequate packaged server.
 
The major problem was the server packaging not the processor. They tested the Elbrus 8S, need to wait the test with Elbrus 16S and an adequate packaged server.
So why not just replicate the existing setup to see if the Elbrus can handle it? I know this question is really out of our hands, but the testing methodology is just flabbergasting. Not even a fair comparison.
 
As far I understand by the Sberbank words, a side of the lower than Xeon performance - albeit considerable higher than expected, this is the actual motherboard/server design that is really unsuitable for them. This is MCST job to provide the correct thing and support, including because other potential clients should bring the same or similar issues.

This is important to note MCST basically is a living Soviet research institute, they don't have a lot of clue of what is used in industry. Also, this test was not for PR but to address what is necessary for potential adoption.

This is also interesting to point out Sberbank is far from being just a bank these days, they are heavily invested "technology" things, in some terms rivaling with Yandex.
 
As far I understand by the Sberbank words, a side of the lower than Xeon performance - albeit considerable higher than expected, this is the actual motherboard/server design that is really unsuitable for them. This is MCST job to provide the correct thing and support, including because other potential clients should bring the same or similar issues.

This is important to note MCST basically is a living Soviet research institute, they don't have a lot of clue of what is used in industry. Also, this test was not for PR but to address what is necessary for potential adoption.

This is also interesting to point out Sberbank is far from being just a bank these days, they are heavily invested "technology" things, in some terms rivaling with Yandex.
well it seems, that even if they don't have a knowledge, today what we have of documentation, on the internet and people available to work, and today the Russian scientists, don't have the mentality of the soviet union.
 
Back
Top