Russian Electronics: Semiconductor and Processors

sepheronx wrote:I was reading comments section on sdelanounas.ru on the computer using the baikal cup and good God are people negative and making crap up. They say all components are Chinese and that nothing is Russian and import substitution is all security and none for commercial, etc etc etc. I think the website got jacked to because I don't recognize any of the posters.

sepheronx wrote:I thought Mikron is already developing 65nm processors as they showcased a while back 65nm on 200mm wafer?

kvs wrote:The bottom line for fabrication is that it is economically justified to contract out instead of building $4 billion fabrication plants.    The existing
Russian fabrication process level (90 nm transitioning into 65 nm) is sufficient for Russian military needs.   For consumer needs TMSC can
serve the function of manufacturing just fine.    The opinions of these retards are not worth the time of day.

Vann7 wrote:

Kimppis wrote:It seems that Russia isn't doing too well in supercomputers:

http://www.unz.com/akarlin/moores-law-for-supercomputers/

Any future plans? Are they only using domestic chips or why do they rank so low? Isn't that a major issue?

First and foremost..

The term "Super computer" is very misleading. It can mean anything .Many of those servers use hardware a decade old..the magic of the "super thing" comes by spending millions dollars in buying thousands of computers and organizing them in a network ,to work as a team. can you imagine a computer how obsolete becomes after 10 years?  All those servers contest about who is the faster is just a benchmark of the speed of a combination of many hundreds or thousands of computers together calculating a mathematical operation. And the speed at which those "super computers" , that will be better called Super networks. Is highly dependent on how much money you invest on it.

Means that if Russia wanted in this year to take the #1 position.. all they need to create a network bigger than the one China have..spend more money than them and configure /organize/Program the servers so no computer is idle doing nothing and that every server is always busy..whenever running any test.

That said. Any nation with money can create a super computer and be the fastest. it all depends of how much $$ money you invest in the servers. Those "Super computers" use ,are very limited. Usually they used to simulate Weather in the world.. or simulate earthquakes..
no big deal.. the most important task could be sharing data.. like Facebook for example need to have Super Computers to actualize billions of tons of data that people enter in the world.

But if you do a test in real time graphics performance ,something like video games.
then even a Playstation 4 will beat the hell of those top 500 super computers in running
graphics in real time. That means ,that top 500 super computers it means next to nothing. that X country managed to organize lots of servers to do run one useless test very fast.

Perhaps those supercomputers will be very fast running mathlab or autocad. but if you pick for example a program like adobe video or Photoshop,it will lose the "super part" and become a super crap , its performance will be no better than your top of the line single personal PC.

So the best name for those "super computers.. will be something like Super Limited Super Servers. Because contrary to top of the line Personal Computers that people can buy , that can do everything. Those "SUper" Servers are only designed with one particular software or application in mind. either to Run Autocad/or Matlab very fast.
or Transfer millions or billions of server files very fast.. something like transfering the personal medical record of a million of people from one server to another.. etc.

So there is nothing to look at the "super computers". Its a meaningless benchmark.
Because do not measure new technology. it only measure Money invested ,organization of servers and good configuration to perform one or two task very fast. and nothing else.

What will be a real Test.. and really interesting .is to compare future Elbrus processors vs Intel I7 processors. In the most used business applications and video games. China design no computers , Russia does. BUilding a super computer is all about going shopping and spending hundred of millions dollars in Intel or Amd or IBM servers. and organizing,configuring them well ,for better performance in a limited applications.

All Said..
Why would Russia need a "super computer"?
it will not help Pak-FA to be better.. neither Armata. Neither will help Russia space program.
Probably to predict earthquakes "better" but not even Japan had any success with that . the most practical use of them will be for Civil Engineer , to calculate very huge structures.
SOmething like the China Super bridge for example had to be simulated there. But it could have been done in any other super server too in the top 500 or top 1000. but less fast.

A real super computer for me.. will be something totally new.. lets say quantum computers ,that
do mathematical computations manipulating atoms. and that such technology is fully developed for consumer and business daily use and that runs all today application
but a hundred of times faster.  That will be really BIG.  

Since the silicon based computers are with us..since when? since the 60's..
Anyone that develops a computer that its circuits no longer use semiconductors /silicon and manipulate atoms this time and instead of electricity use light,that will be HUGEEEEEE as big
as landing on mars today ,building a base while filming it and then return a year later safely .

This Elbrus Computer Russia is building is much more prestigious and more big deal
than a going shopping of thousand american processors and organizing them ,to create the
fastest Network in the world. This is because Russia is developing new Technology that could replace Americans one ,while China is not.

It looks like Russia does not need to
simulate its nuclear devices like the USA. One can laugh and dick stroke about how Russians are inferior, but
they have always relied on mathematical approaches (e.g. asymptotic perturbation expansions) and pencil
and paper instead of a big numerical box. They have clearly done quite well with their approach.

Correct.. Russia have no need for being in the top 10 or top 100 supercomputers. To build
a super server just for the fun of it and bragging rights is total silly . Yeah physical simulations,
weather ,counting stars in the galaxy?      Japan used their super computers to predict
earthquakes and so far no one will believe they have been successful in that. The best use
i see for Super Computer is in Entertainment industry    , to render those nice shots of
3d movies in no time. but also for engineering of super giant structures.. like very long bridges
or buildings..and very quickly do double check by hand and pencil and by computer of everything.  Russia dont need a super computer..  but they do need to d comesign completely NEW COmputers never done before ,for import substition and for security ,to replace American technology in Russia.

the next article will illustrate more about super computers.. how they are used .

But what the heck can you do with a super computer?

The short answer is: Not much. Or at least, nothing that mere mortals like you or I usually concern ourselves with.  

http://www.extremetech.com/extreme/122159-what-can-you-do-with-a-supercomputer

But is not so bad.. Supercomputers is used for research by scientist. The problem however
comes ,for the very fact that the super speed will be totally useless even in the few things
a super computer can be used.. if the program used is not done right. So if for example you
have 1 million computers trying to solve a mathematical problem.. but one programmer by mistake enters the wrong value or formula.. then the entire work will be wasted.  Im really
skeptic about the benefits of them. For the fact that computers do not work with real time
information. things that happens in the now. computer data is always recorded ,stored and is old. So if you for example enter in a super computer all the traffic accidents in the world ,and how they happened. the computer will be still unable to predict future accidents with total accuracy. all said to study the behavior of nature or of the human body and to try to predict the future by looking at old data , is doomed to failure. because such things are random and do not follow any script or predictable pattern. Japan have failed in the use of their Supercomputers to
predict earthquakes for examples.. and Super computers will also fail to predict illness or find cure for anything. the right place for Super Computers should not be research.. or to "discover anything" in my opinion.. but simply to speed things , that human can do ,but will take them much more time.. It also can be used to organize endless data so that you can more easier study it and find exactly what you want.

So in summary Computers/super computers,are useless to discover anything new.
Because they operate with the data that you enter ,so do not reason..
What Super Computers can do , is help Organize Mega Tons of Information that you enter
,so that it will help you see things.. to better study your own information ,that is to make it easier to visualize things. but also to speed calculations in mathematics. and display graphics too.  But Super Computers will never be able to discover anything new or create new science.
So this is the reason Why Russia don't need a "super computer" in the top positions because there is no use for it.

Russian and Japanese scientists have developed graphene-based transistor and demonstrated with modeling the feasibility of its ultralow-power operation, the Moscow Institute of Physics and Technology (MIPT) said on Monday.

The most important consequence of reduced power consumption is the increase in performance which could be as high as two orders of magnitude. These record values are possible due to the unusual dependence of electron energy on momentum in graphene bilayer, which resembles a Mexican hat.

sepheronx wrote:

Singular_trafo wrote:Interesting qestion, if the S-500 will be capable to intercept high level ICBMs and nuclear warhead , then it must has radiation resistant electronics.

By my best knowledge russia doesn't has capacity to manufacture these circuits.
They will use Chinese electronics in the interceptor rockets?

Additionaly, the S-X00 radar can handle ECM blindness?

Russia's military microelectronics are made in Zelenograd. Micron plant and Angstrom-T.

sepheronx wrote:China uses basic Motorola processors for mil in radiation environment. What do you think Russia been using for years? They have a wide selection. Elbrus 2C is designed for radar use. Hence need anti radiation shielding. And even then, it is all in special housing.

I suggest looking it up. Maybe do a bit more research.

These are a good start:
http://igorrgroup.blogspot.ca/2009/11/russian-chip-makers-part-1.html
and
http://igorrgroup.blogspot.ca/2009/11/russian-chip-makers-part-2.html

producers:
http://multicore.ru/
http://www.mikron.ru/en/
http://www.module.ru/

Of course there are probably subcomponents that need to be imported. And as you mentioned, they just started recently on MRAM so that was another one they had to import in the past.

kvs wrote:The Russian military has been using custom ASICs for years. CPUs are overkill in many applications and having custom
logic can do a better and faster job than software emulation on a generic CPU. This aspect is totally lost on many
"experts".

http://www.mri-progress.ru/en/about/info/

http://cordis.europa.eu/docs/projects/cnect/2/247992/080/deliverables/001-Annex6D21SEMIDEC.pdf

sepheronx wrote:

And China isn't major producer either. Actually, if looking back only a few years, US has been using their Intel i960 processors for such applications. What module and Elvees supplies are modern chips of both MIPS and now ARM as well. Their use in satellites are obvious indication on radioactive shielding (not all of it are onboard based. Shielding is and has been done on housing) due to massive radiation levels in space.

Singular_trafo wrote:Interesting qestion, if the S-500 will be capable to intercept high level ICBMs and nuclear warhead , then it must has radiation resistant electronics.

By my best knowledge russia doesn't has capacity to manufacture these circuits.
They will use Chinese electronics in the interceptor rockets?

Additionaly, the S-X00 radar can handle ECM blindness?

Vann7 wrote:

Singular_trafo wrote:Interesting qestion, if the S-500 will be capable to intercept high level ICBMs and nuclear warhead , then it must has radiation resistant electronics.

By my best knowledge russia doesn't has capacity to manufacture these circuits.
They will use Chinese electronics in the interceptor rockets?

Additionaly, the S-X00 radar can handle ECM blindness?

Dude , i really think you are seriously misinformed. wikipedia is not really the best
source of information ,just get that right for once. and Russia was first in space, first in the Moon ,Mars and Venus a decade or two earlier before than NASA in most of them ,and in Venus NASA not even have been there, and the radiation in open space is the most brutal you can get ,not mentioned either the below zero ,-270.45 Celsius. that any probe needs to survive there.

Lunokhod the Russian Rover was in the moon in 1970 for more than a year , taking photos.
and China was not remotely a power in those time ,even less had any technology. So contrary
to your believe Russia do have plenty of experience dealing with Radiation, even more than
your wikipedia reality claims.

Institute of electronic control machines them. I. S. Bruk (INEUM, part of United Instrument manufacturing Corporation) in conjunction with the MCST, JSC and OOO TESIS successfully completed the first phase of the transfer to platform Elbrus software complex computational Aero - and hydrodynamics is FlowVision  by creating a working prototype of a fully Russian hardware-software complex for performing industrial engineering calculations.
Industrially produced serial processors and the availability of tools for developers on the platform Elbrus, including the compilation tool gave an opportunity to the developers of the company TESIS start initiative work on porting the x86 version of FlowVision on a platform Elbrus. Client-server modular architecture is Studied, written in C++, allowed in the shortest possible time to perform the porting of the core modules. Technical Director TESIS Andrey Aksenov said: "we have not had any major difficulties and problems when migrating a version of FlowVision on a platform Elbrus, and all technical issues were resolved in a fast and qualified support programmers AO MCST".
Currently the software package FlowVision allows to use server four-processor Elbrus-4.4 for calculations and workstation Elbrus‑401 for visualization and results analysis. Due to the high degree of parallelism is Studied full-scale launch of numerical simulation possible on a compute cluster of servers Elbrus-4.4. In particular, a cluster of servers Elbrus-4.4 with a total peak performance of 13.8 teraflops with up to 64 servers, was developed in INEUM by order of the Ministry of industry and trade.

Thanks to the joint efforts of specialists INEUM, MCST and TESIS, an important step towards the creation of a fully Russian software and hardware systems for the solution of engineering tasks throughout the product life cycle. This work is due to the increased requirements on information security and demand from enterprises of the military-industrial complex, space industry, aviation and shipbuilding industries and enterprises of Rosatom.
The hardware and software complex FlowVision-Elbrus" for solving problems of computational hydro - and aerodynamics has the following characteristics:
comparable in volume and quality of opportunities and global level, and in some features superior to many foreign products;
ready to use at the enterprises of defense industry and other industries, came under Western sanctions;
the ruble has competitive prices irrespective of currency fluctuations.
Software complex computational Aero - and hydrodynamics is Studied on the platform of Elbrus was first presented at the forum "the Digital industry industrial Russia" (CIPROFLOXACIN) in Innopolis June 7-10, 2016.
It should be noted that the decision on transfer of the software complex FlowVision on the platform Elbrus was made by companies MCST and TESIS on the National supercomputer ForumNSCP-2015, in which both companies took part (for more details see: MCST and TESIS). The forum is held in the ISS named after A. K. aylamazyan ran with the support of the FANO Russia and Department of nanotechnologies and information technologies of RAS, dedicated to the creation and practical application of supercomputer technologies, promotes development of cooperation of Russian companies and enterprises on the subject. Also to work with the platform Elbrus on a proactive basis joined the team of the chair of "Electronic computing machines and systems"from the Volgograd Polytechnic University, which started research the performance of the Elbrus architecture for different computational algorithms.
The forum NSCF-2016, which will begin its work on 29 November 2016, MCST and the TESIS will submit a detailed report on the work done, as well as demonstrate the work of local hardware-software complex "FlowVision-Elbrus" for the execution of industrial engineering calculations.
Institute of electronic control machines them. I. S. Bruk included in "Combined instrument Corporation (OPK, is included in the state Corporation "rostec"). Development servers Elbrus 4.4 was supported by the Ministry of industry and trade of Russia.

sepheronx wrote:"Angstrom" has developed a unique space transistors