Russian Electronics: Semiconductor and Processors

The other thing too, is the software behind it all

So I would be interested to know when they will release Elbrus 2SM which is supposed to be commercial.

Кремниевый транзисторный ключ, используемый для модуляторов питания усилителей мощности на основе GaAs-транзисторов, является функциональным аналогом интеллектуального ключа – BTS640 фирмы Infineon Technologies AS (Германия).


Для кремниевого транзисторного ключа, используемого для модуляторов питания усилителей мощности на основе GaN-транзисторов, прямые отечественные и зарубежные аналоги отсутствуют.

According to the newspaper "Kommersant", in 2015 in Russia will create a range of domestic microprocessors Baikal topology 28 nm

The first products in the line should be eight-core processors Baikal M and M / S topology of 28 nm and 2 GHz for use in personal computers and mikroserverah. Processors will be built on a 64-bit kernel Cortex A-57 British company ARM. As the operating system for all solutions to be used free software Linux.

Issue eight-processor Baikal is scheduled for early 2015, and in late 2016 the Ministry of Industry will release even more powerful, 16-core server processor topology 16 nm.

Russian Patriot wrote:http://en.itar-tass.com/economy/736804 new computer chips to replace American Intel and AMD

sepheronx wrote:On ARM? Strange. They mention topology of 28nm and 8 core, which is what Elbrus 8C will be. Which isnt arm but CISC.

Although, obtaining ARM license is cheap and apparently production us cheap too, so I can see the feesibility. Add in that AMD is hoping to release their own ARM server grade Opteron processors, so I suppose there are server grade software already available that takes advantage of the architecture of the processor.

Rosteh has the money and technology, while T-Platforms has knowhow and what is needed to make HPC grade computers. So I look forward to its development. As well, if it is true, then they will be a fabless facility and contract out whatever FAB that will do the 28nm topology.

Edit: add to that, they should get China or Chinese company involved in its development as well as opening a FAB or contracting a FAB to build ones in China to both reduce costs and get other players/investors involved (which will translate to more companies building accessories for these processors and more software developers. Which is what is holding Elbrus processors back. Not much in terms of software besides some in house built ones, that takes advantage of Elbrus arch.

Strizh wrote:

sepheronx wrote:On ARM? Strange. They mention topology of 28nm and 8 core, which is what Elbrus 8C will be. Which isnt arm but CISC.

Although, obtaining ARM license is cheap and apparently production us cheap too, so I can see the feesibility. Add in that AMD is hoping to release their own ARM server grade Opteron processors, so I suppose there are server grade software already available that takes advantage of the architecture of the processor.

Rosteh has the money and technology, while T-Platforms has knowhow and what is needed to make HPC grade computers. So I look forward to its development. As well, if it is true, then they will be a fabless facility and contract out whatever FAB that will do the 28nm topology.

Edit: add to that, they should get China or Chinese company involved in its development as well as opening a FAB or contracting a FAB to build ones in China to both reduce costs and get other players/investors involved (which will translate to more companies building accessories for these processors and more software developers. Which is what is holding Elbrus processors back. Not much in terms of software besides some in house built ones, that takes advantage of Elbrus arch.

Mhh I still very much dislike that the CPU will be produced outside of Russia! Russia should build at least one FAB on it's own territory which is able to handle 28-16nm.

vK_man wrote:

Strizh wrote:

sepheronx wrote:On ARM? Strange. They mention topology of 28nm and 8 core, which is what Elbrus 8C will be. Which isnt arm but CISC.

Although, obtaining ARM license is cheap and apparently production us cheap too, so I can see the feesibility. Add in that AMD is hoping to release their own ARM server grade Opteron processors, so I suppose there are server grade software already available that takes advantage of the architecture of the processor.

Rosteh has the money and technology, while T-Platforms has knowhow and what is needed to make HPC grade computers. So I look forward to its development. As well, if it is true, then they will be a fabless facility and contract out whatever FAB that will do the 28nm topology.

Edit: add to that, they should get China or Chinese company involved in its development as well as opening a FAB or contracting a FAB to build ones in China to both reduce costs and get other players/investors involved (which will translate to more companies building accessories for these processors and more software developers. Which is what is holding Elbrus processors back. Not much in terms of software besides some in house built ones, that takes advantage of Elbrus arch.

Mhh I still very much dislike that the CPU will be produced outside of Russia! Russia should build at least one FAB on it's own territory which is able to handle 28-16nm.

where?

Viktor wrote:@sepheronx - whats going on here

2014In Russia, organize production 3D-microsystems

sepheronx wrote:

vK_man wrote:

Strizh wrote:

sepheronx wrote:On ARM? Strange. They mention topology of 28nm and 8 core, which is what Elbrus 8C will be. Which isnt arm but CISC.

Although, obtaining ARM license is cheap and apparently production us cheap too, so I can see the feesibility. Add in that AMD is hoping to release their own ARM server grade Opteron processors, so I suppose there are server grade software already available that takes advantage of the architecture of the processor.

Rosteh has the money and technology, while T-Platforms has knowhow and what is needed to make HPC grade computers. So I look forward to its development. As well, if it is true, then they will be a fabless facility and contract out whatever FAB that will do the 28nm topology.

Edit: add to that, they should get China or Chinese company involved in its development as well as opening a FAB or contracting a FAB to build ones in China to both reduce costs and get other players/investors involved (which will translate to more companies building accessories for these processors and more software developers. Which is what is holding Elbrus processors back. Not much in terms of software besides some in house built ones, that takes advantage of Elbrus arch.

Mhh I still very much dislike that the CPU will be produced outside of Russia! Russia should build at least one FAB on it's own territory which is able to handle 28-16nm.

where?

Taiwan.  Elbrus 8C will be built in Taiwan as there is no 28nm FAB in Russia.  Only 65 and higher.  Hopefully they will spend the $1+B in a FAB in Russia.  But right now, it is much easier and cheaper to do it outside.

Viktor wrote:@sepheronx - whats going on here

2014In Russia, organize production 3D-microsystems

Sounds like a way to build modern era and future microscope, lenses and other technology.  I know very little in 3D microsystems.  I will have to do more investigations on it.

vK_man wrote:

sepheronx wrote:

vK_man wrote:

Strizh wrote:

sepheronx wrote:On ARM? Strange. They mention topology of 28nm and 8 core, which is what Elbrus 8C will be. Which isnt arm but CISC.

Although, obtaining ARM license is cheap and apparently production us cheap too, so I can see the feesibility. Add in that AMD is hoping to release their own ARM server grade Opteron processors, so I suppose there are server grade software already available that takes advantage of the architecture of the processor.

Rosteh has the money and technology, while T-Platforms has knowhow and what is needed to make HPC grade computers. So I look forward to its development. As well, if it is true, then they will be a fabless facility and contract out whatever FAB that will do the 28nm topology.

Edit: add to that, they should get China or Chinese company involved in its development as well as opening a FAB or contracting a FAB to build ones in China to both reduce costs and get other players/investors involved (which will translate to more companies building accessories for these processors and more software developers. Which is what is holding Elbrus processors back. Not much in terms of software besides some in house built ones, that takes advantage of Elbrus arch.

Mhh I still very much dislike that the CPU will be produced outside of Russia! Russia should build at least one FAB on it's own territory which is able to handle 28-16nm.

where?

Taiwan.  Elbrus 8C will be built in Taiwan as there is no 28nm FAB in Russia.  Only 65 and higher.  Hopefully they will spend the $1+B in a FAB in Russia.  But right now, it is much easier and cheaper to do it outside.

Viktor wrote:@sepheronx - whats going on here

2014In Russia, organize production 3D-microsystems

Sounds like a way to build modern era and future microscope, lenses and other technology.  I know very little in 3D microsystems.  I will have to do more investigations on it.

Russia should try for 3D printing CPU's . 3D printing is more flexible than conventional method of production.

ZAO "MCST" launched the pilot batch universal microprocessors Elbrus-8C . Rated operational frequency chip - 1.3 GHz, manufacturing technology - 28 nm, the computational power of 250 gigaflops. Getting ready-made samples of chips is expected in October 2014.
Elbrus microprocessor-8C - completely Russian development. Crystal microprocessor designed for 28 nm technology, has 8 cores with improved 64-bit architecture Elbrus third generation cache level 2, totaling 4 megabytes and third level of 16 megabytes.
Microprocessor based Elbrus-8C planned mass production servers, workstations and other computer equipment intended for use in public institutions and business structures, placing increased demands on information security, as well as for use in high-performance computing, signal processing, telecommunications. Engineering samples of 4-processor servers based on the Elbrus-8C 1 teraflops of performance will be made at the end of 2014.
Domestic architecture developed Elbrus in Russia and has some unique features. These include:
ability to perform on each core to 25 operations in a single clock cycle, which provides high performance at moderate clock frequency;
dynamic binary translation technology, to ensure the effective execution of applications and operating systems, distributed in binary codes x86, including multi-threaded;
mode support secure computing hardware with a special control structure integrity of memory, which allows a high level of information security using its software systems.
The operating system platform is OS Elbrus "Elbrus", built on the basis of kernel Linux. Programming system platform supports languages ​​C, C + +, Java, Fortran 77, Fortran 90.

On July 3 the territory of Technopolis "Moscow" started production line of one of the most intensive and central components bezmasochnyh lithographers Dutch company Mapper Lithography - elements of electron optics MEMS (microelectromechanical systems).

The company, one of whose shareholders is RUSNANO has invested in a new production at Technopolis "Moscow" about 1 billion rubles.

"Mapper Lithography" more than 10 years developing the technology of electron-beam lithography bezmasochnoy. Lithography machines are used in the manufacture of semiconductor devices, integrated circuits, as well as some of superconducting nanostructures.
It is planned that a running production line fully cover the needs of the company in the elements of electron optics for lithographic machines.

Moscow factory area of ​​two thousand square meters. Of these, half are clean rooms, which, according to the measurements correspond cleanliness class according to ISO 6.
The production line is fully equipped with modern high-tech equipment.

The plant will operate a team of 30 Russian specialists trained in the Netherlands. It is noteworthy that the work in this ultra-modern production returned to Russia and some engineers previously left the country and in the West have been successful in leading international companies.

Lithography is the central process of production of integrated circuits, semiconductor devices, as well as some of superconducting nanostructures. Industry standard today - optical immersion lithography, which provides a resolution of about 32 nanometers. A number of additional technology allows to increase the resolution of optical lithography to 22 nanometers, in a tangible way the economy worsens production.
Alternative to optical lithography is bezmasochnaya lithography based on beam technology (Electron-beam lithography, E-beam). Mapper Lithography more than 10 years developing the technology of electron-beam lithography and bezmasochnoy already has several industrial prototypes lithographer, which were purchased and were tested by the leading players in the industry - the Taiwanese company TSMC and Research Institute of Microelectronics CEA-Leti (France). At the moment, the equipment allows the company to reach a resolution in the 22 nanometer.
Russia produces a key component of lithographic systems Mapper - electron optics based on microelectromechanical systems (MEMS). Innovative solutions Mapper and their main difference from competitors is to use more than one electron beam, as was done earlier, and 13,000 beams simultaneously, which dramatically increases system performance. Elements of electron optics manufactured in Russia, just one ray is converted to 13,000 and run each such beam separately.
It is assumed that the plant will be constructed in Moscow three types of electron optics. The most simple - spacers - are used to separate the elements of electron optics. The following structures of complexity - silicon electron lenses are designed for focusing and collimation of electron beams. Their production will begin by the end of this year. Release of the most difficult elements - containing electronics control electrodes - is scheduled to begin by the end of 2015.
Lithography equipment market is highly consolidated and, in fact, represented by three major producers: ASML, Nikon and Canon. Market size is about $ 6 billion, which corresponds to sales of several hundred cars a year.
By the time the design capacity, the Russian plant Mapper will produce sets of electron optics for 20 cars per year.

Products Mapper is focused on two market niches. First, lithographers company designed for small and medium sized little-chip manufacturers. No need to book a mask (one of the most expensive elements in microelectronics manufacturing) makes cost-effective for companies such issue even minimal batches of chips. Secondly, products Mapper addressed and large companies interested in the speedy promotion of new products on the market. At the design stage of the chip manufacturer forced to order several masks. In case of error, the mask has to remodel, resulting in time and cost.
Mapper allows equipment at the design stage to test a virtually unlimited number of design options without having to create multiple lithographic masks, which significantly reduces the cost and faster time to market the new product.
Video about the technology in the language of a potential enemy: https://www.youtube.com/wa...mbedded&v=OQBcDbhw-0Y
Article in Russian with explanations: http://habrahabr.ru/post/213379/

In early 2014, the Moscow Center of SPARC Technologies (MCST) launched a test production of Elbrus-8C eight-core processors (topology 28 nm, proprietary 64-bit architecture, clock speed 3.3 GHz, computation capacity 250 GFlops). Its development began in 2011 on a commission from the Industry and Trade Ministry, which has invested $24.4 million in the project. The manufacturers promise to produce around 300 Elbrus-8C microchips by the end of 2014. In future, MCST intends to produce servers, workstations and telecom equipment based on its microchips. Elbrus-8C is the first eight-core processor to be made by a Russian company. Several months ago the Industry and Trade Ministry and the Defense Ministry raised the issue of substituting imported electronics, in particular Intel and AMD processors, in connection with the latest U.S. and EU sanctions and the political confrontation over Ukraine.

MCST expects that the list of potential customers for the new microchip will include major Russian manufacturers of computer equipment such as Depo Computers, Kraftway, and Akvarius. In 2012, Kraftway launched the production of all-in-one PCs based on MCST's previous-generation processor, Elbrus-2C. However, due to the small scale of production, the price of each unit exceeded $3,000, which prevented it from making it to the general market. This time round, according to MCST's Konstantin Trushkin, the company hopes to sell its Elbrus-8C-based products to universities and state structures in the expectation that this will allow it to expand production and thus reduce the price.

These plans are supported by the Industry and Trade Ministry's statistics: according to it, state companies and organizations purchase about 700,000 personal devices every year worth a total of $500 million as well as 300,000 servers worth $800 million, with the overall market amounting to $3.5 billion. "Elbrus-8C microchips will be made abroad. In Russia, there is no technological capacity for that. Our microprocessors will be manufactured by Taiwan's TSMC," Trushkin said in an interview with the Kommersant newspaper.

In addition to Elbrus microchips, MCST is working on a more powerful microchip: In May 2014, it signed a contract with the Industry and Trade Ministry for the development of a 64-bit processor, El-16C, with a process of under 20 nm, and eight to 16 cores. The chip is expected to be developed by November 2018 and is proposed to be used for the needs of the Russian Ministry of Defense. Uncertain prospects Experts are not convinced that the new microprocessor has much potential. According to Oleg Kolenchenko, a hardware reviewer at the specialist magazine Ferra, Elbrus microprocessors are unlikely to be in a position to compete with Intel, AMD and IBM. "However, one should not underestimate the significance of work on new microprocessors. The release of the Elbrus-8C is a very important step forward. All the more so since its architecture features several very interesting technologies, especially those to do with security, and the microprocessor performs its defense sector tasks, which it was commissioned for in the first place, very confidently," Kolenchenko

Intel representatives do not yet see any threats to its operations in Russia either. "We are an American company and, of course, comply with the U.S. authorities' decisions. However, judging by the information about the sanctions that is in the public domain at present, Intel will not have to adjust its operations in Russia in any way," Intel's regional director for the CIS Dmitry Konash has told RBTH. "We do not supply anything directly to the companies on the sanctions list or controlled by individuals who are on the sanctions list. We have a number of clients in Russia who we work with directly, on technology consulting projects." In addition to MCST, there are several more Russian companies manufacturing computers and microchips for them: Rover Computers, best known for its RoverBook laptops; DEPO Computers, which focuses on the production of desktop computers and servers; and Desten, which manufactures home and specialist office PCs.

sepheronx wrote:
That aside, after looking at the only review of the Elbrus 4C through CNews.ru, I can safely say that the processor shows a lot of potential.  The fact that it came close to competing against a Core i3 and i5, with having little to no real CISC processor development experience before, besides working on the Elbrus E2K core, has shown Russia is more than capable of producing a chip that can easily compete with the big players.

...

Baikal processors through Baikal electronics, subsidiary of T-Platforms may have a higher chance of being a direct competitor (and it will be good for MCST as they will have internal competition), due to the fact that T-Platforms (a major HPC making company) will be using these processors in the development of their cluster systems, which they seem to have a bigger access to the market than MCST as they have super computers built even in the USA.  So in this case, T-Platforms Baikal processors (whenever they will end up showing up for sale), could very well be the basic market consumer processor, especially since ARM processors are making a huge push into the consumer market (good for a RISC processor).

kvs wrote:

sepheronx wrote:
That aside, after looking at the only review of the Elbrus 4C through CNews.ru, I can safely say that the processor shows a lot of potential.  The fact that it came close to competing against a Core i3 and i5, with having little to no real CISC processor development experience before, besides working on the Elbrus E2K core, has shown Russia is more than capable of producing a chip that can easily compete with the big players.

...

Baikal processors through Baikal electronics, subsidiary of T-Platforms may have a higher chance of being a direct competitor (and it will be good for MCST as they will have internal competition), due to the fact that T-Platforms (a major HPC making company) will be using these processors in the development of their cluster systems, which they seem to have a bigger access to the market than MCST as they have super computers built even in the USA.  So in this case, T-Platforms Baikal processors (whenever they will end up showing up for sale), could very well be the basic market consumer processor, especially since ARM processors are making a huge push into the consumer market (good for a RISC processor).

The Elbrus CPU line is a competitor for Xeons and Opterons and should be deployed in Russia for scientific and engineering computing.  T-Platforms should adopt this CPU instead of relying on Intel or AMD parts.  Unfortunately, the Elbrus CPUs take a large performance hit when doing 64 bit floating point operations compared to 32 bit operations.   This is the price of having a low power design with fewer transistors.   MCST should release a variant that does not have this limitation and this does not require a ground-up redesign.  

The Elbrus design has demonstrated that VLIW can work just fine and Russia has managed this feat.   Intel tried VLIW with the Itanium line but has failed.

sepheronx wrote:

kvs wrote:

sepheronx wrote:
That aside, after looking at the only review of the Elbrus 4C through CNews.ru, I can safely say that the processor shows a lot of potential.  The fact that it came close to competing against a Core i3 and i5, with having little to no real CISC processor development experience before, besides working on the Elbrus E2K core, has shown Russia is more than capable of producing a chip that can easily compete with the big players.

...

Baikal processors through Baikal electronics, subsidiary of T-Platforms may have a higher chance of being a direct competitor (and it will be good for MCST as they will have internal competition), due to the fact that T-Platforms (a major HPC making company) will be using these processors in the development of their cluster systems, which they seem to have a bigger access to the market than MCST as they have super computers built even in the USA.  So in this case, T-Platforms Baikal processors (whenever they will end up showing up for sale), could very well be the basic market consumer processor, especially since ARM processors are making a huge push into the consumer market (good for a RISC processor).

The Elbrus CPU line is a competitor for Xeons and Opterons and should be deployed in Russia for scientific and engineering computing.  T-Platforms should adopt this CPU instead of relying on Intel or AMD parts.  Unfortunately, the Elbrus CPUs take a large performance hit when doing 64 bit floating point operations compared to 32 bit operations.   This is the price of having a low power design with fewer transistors.   MCST should release a variant that does not have this limitation and this does not require a ground-up redesign.  

The Elbrus design has demonstrated that VLIW can work just fine and Russia has managed this feat.   Intel tried VLIW with the Itanium line but has failed.

Exactly.  Problem with 64bit is that VLIW isn't native x86 thus it has to have a separate instruction for it.  And thus, that is where the performance hit is (theoretically) compared to lets say Intel or AMD's since both of those are CISC (I had it wrong the whole time, thought Elbrus was CISC and not VLIW).  That being said, I agree, they can indeed increase transistor counts significantly.  I guess we will not know fully until when Elbrus 8C is shown.  That said, am looking forward to see what it will be, as well, what the future holds after that.  I know they want to move as well onto APU's.  Now Baikal electronics is moving into RISC with its future ARM processor and any other processor afterwards.

kvs wrote:

sepheronx wrote:

kvs wrote:

sepheronx wrote:
That aside, after looking at the only review of the Elbrus 4C through CNews.ru, I can safely say that the processor shows a lot of potential.  The fact that it came close to competing against a Core i3 and i5, with having little to no real CISC processor development experience before, besides working on the Elbrus E2K core, has shown Russia is more than capable of producing a chip that can easily compete with the big players.

...

Baikal processors through Baikal electronics, subsidiary of T-Platforms may have a higher chance of being a direct competitor (and it will be good for MCST as they will have internal competition), due to the fact that T-Platforms (a major HPC making company) will be using these processors in the development of their cluster systems, which they seem to have a bigger access to the market than MCST as they have super computers built even in the USA.  So in this case, T-Platforms Baikal processors (whenever they will end up showing up for sale), could very well be the basic market consumer processor, especially since ARM processors are making a huge push into the consumer market (good for a RISC processor).

The Elbrus CPU line is a competitor for Xeons and Opterons and should be deployed in Russia for scientific and engineering computing.  T-Platforms should adopt this CPU instead of relying on Intel or AMD parts.  Unfortunately, the Elbrus CPUs take a large performance hit when doing 64 bit floating point operations compared to 32 bit operations.   This is the price of having a low power design with fewer transistors.   MCST should release a variant that does not have this limitation and this does not require a ground-up redesign.  

The Elbrus design has demonstrated that VLIW can work just fine and Russia has managed this feat.   Intel tried VLIW with the Itanium line but has failed.

Exactly.  Problem with 64bit is that VLIW isn't native x86 thus it has to have a separate instruction for it.  And thus, that is where the performance hit is (theoretically) compared to lets say Intel or AMD's since both of those are CISC (I had it wrong the whole time, thought Elbrus was CISC and not VLIW).  That being said, I agree, they can indeed increase transistor counts significantly.  I guess we will not know fully until when Elbrus 8C is shown.  That said, am looking forward to see what it will be, as well, what the future holds after that.  I know they want to move as well onto APU's.  Now Baikal electronics is moving into RISC with its future ARM processor and any other processor afterwards.

I forgot to clarify the point about software.  For science and engineering what one needs is a good Fortran 90 compiler and there are not that many useful off the shelf software packages.    There are more commercial packages for engineering, but even there it is typical to resort to custom code.   So the CISC vs. VLIW architecture differences are not a problem.   Also, these days you cannot get far without a large amount of CPUs operating in parallel via MPI and to a lesser extend OpenMP.   Russia needs to develop good clusters based on the Elbrus.  

The Baikal CPU is based on the ARM architecture and is good for integer heavy business computing.  

It was good to hear that Mikron has started producing the Elbrus-2S in Russia and has upgraded their equipment to produce 65 nm parts.  The fabrication equipment was rated for 90 nm production initially.   So Russia is developing, slowly, a proper microelectronics industry.

sepheronx wrote:Well, i know sputnik (developer of mini satellitrs, private company in Russia) partnered up with Microclet to produce new processors for space.