Chinese semiconductor industry

Status
Not open for further replies.

supercat

Major
Chinese company launches the country's first homegrown 7nm GPGPU chip
China has made a major breakthrough in leading-edge 7nm computing chips that could rival products from global industry leaders like U.S. companies Nvidia and AMD.

Shanghai Tianshu Zhixin Semiconductor Co., a domestic chipmaker, announced on Wednesday that it had launched the country's first homegrown 7nm GPGPU chip named "Big Island" or "BI."

GPGPU is short for General-Purpose Graphics Processing Unit, which refers to a graphics processing unit (GPU) that can perform non-specialized computations in addition to its traditional purpose of computation for computer graphics, such as computations typically handled by a Central Processing Unit (CPU). It's an overall faster, high-performance processor that combines CPU and GPU processing power.

GPGPU chips not only feature parallel processing capabilities, but can also have high memory bandwidth. Therefore, they are used to power everything from databases to high-performance computing (HPC) and artificial intelligence (AI).

However, the area has long been dominated by international giants such as Nvidia and AMD.

The launch of BI marks a big step forward for China's design and development of GPU chips, which is expected to break the dominance of the technology by foreign companies, said Zhang Ying, deputy director of Shanghai Municipal Commission of Economy and Information Technology.
Please, Log in or Register to view URLs content!
 

voyager1

Captain
Registered Member
Please, Log in or Register to view URLs content!
In the grand theme of things, a useless company.

As @Appix said, one call from Jake Sullivan (US national security advisor) to TSMC and this company will disappear from planet Earth.

What the heck are the other domestic equipment IC companies doing. Central Gov is showering them with money and resources and they will still need 5 (?) Years for domestic 7 nm. I though this project was a national priority, so why does it still takes so long.

I hope the people there work 24/7/365 all year round with no big holidays (with big compensation of course). This is supposed to be a "Manhattan"-level project where everyone would do everything to achieve the goal.

Also, where is espionage? Why have they still didn't take the design and manufactoring blueprints already. Why they dont have agents inside the most technological advanced ASML and its suppliers?

Big failing from the Intelligence Service and from the Central Gov and from these companies also.

So to sum it up, failings everywhere. Now, I am sure they will succeed but they are too slow. The 2025 timeframe for 7nm is too late, between now and then the US will have a constant guillotine hanging on China's head, ready to be dropped at any moment
 
Last edited:

Oldschool

Junior Member
Registered Member
Newly established company , based from tsinghua University, related to u-precision makes optical diffraction grating , core component in optics.



至格科技官网显示,公司成立于2019年,是一家AR衍射光波导及衍射光栅供应商。该公司孵化自清华大学精密仪器系,依托于清华大学二十余年的光栅领域科研成果进行产业转化,致力于增强现实光学显示模组及衍射光栅的研发、生产和销售。

该公司掌握光栅产业链三大核心技术——光栅设计、光栅母版加工和纳米压印生产。在纳米压印生产方面,已经打通全部生产工艺环节,图形转写率可达95%。

According to the official website of Zhige Technology, the company was established in 2019 and is a supplier of AR diffractive optical waveguides and diffraction gratings. The company was incubated from the Department of Precision Instruments of Tsinghua University. It relies on Tsinghua University's more than 20 years of scientific research in the field of gratings for industrial transformation, and is committed to the research and development, production and sales of augmented reality optical display modules and diffraction gratings. The company masters the three core technologies of the grating industry chain-grating design, grating mastering and nanoimprint production. In terms of nano-imprint production, all production process links have been opened up, and the graphics transfer rate can reach 95%.
 

coolieno99

Junior Member
In the grand theme of things, a useless company.

As @Appix said, one call from Jake Sullivan (US national security advisor) to TSMC and this company will disappear from planet Earth.

What the heck are the other domestic equipment IC companies doing. Central Gov is showering them with money and resources and they will still need 5 (?) Years for domestic 7 nm. I though this project was a national priority, so why does it still takes so long.

I hope the people there work 24/7/365 all year round with no big holidays (with big compensation of course). This is supposed to be a "Manhattan"-level project where everyone would do everything to achieve the goal.

Also, where is espionage? Why have they still didn't take the design and manufactoring blueprints already. Why they dont have agents inside the most technological advanced ASML and its suppliers?

Big failing from the Intelligence Service and from the Central Gov and from these companies also.

So to sum it up, failings everywhere. Now, I am sure they will succeed but they are too slow. The 2025 timeframe for 7nm is too late, between now and then the US will have a constant guillotine hanging on China's head, ready to be dropped at any moment
The drawings and design of ASML litho tools are available at US Patent Office for public viewing. There's no need to spy on ASML. China probably already have the drawings to design a workaround to avoid patent infringements. China herself is one of the biggest producers of patents in the world (see Huawei 5G). This is the main reason China does not want to violate patent rights. From a business tactic scheme, Shanghai Tianshu Zhixin Semiconductor Co should restrict sales of its GPGPU chips within China to avoid drawing attention from NVIDIA and AMD. Wait until China gets its EUV 7nm line up and running and then try to sell GPGPU outside China. From reading posts(WTAN, OldSchool, Foofy, ... etc) from the last couple weeks , China is moving at lightening speed to completely indigenize its wafer fab operation that includes all the support tools (photoresist, testing, ...etc) besides the EUV litho. China already built a working EUV litho engineering demonstrator at 28 nm earlier. Now she has to build a 7 nm production litho tool that can process 100 to 200 wafers/hr and not infringing on existing patents. I would not be surprised if she can do this within 1 or 2 years.

Please, Log in or Register to view URLs content!
 

gelgoog

Lieutenant General
Registered Member
There are just too many things to do. It is not just lithography, although that is the most visible component. You need all the tools, and materials including chemicals and masks, then you need to integrate everything. It is one thing to produce a working sample, quite another to achieve mass production with the quality required for these products. Then, even if you have everything, you need to make a production line with all these components and get it working smoothly and effectively enough to compete with existing solution that have been in the market for at least a decade, from vendors who have been in the market for four decades. This was never going to happen in 2 years. It takes 2 years just to build a fab and that is assuming you already had all the components on hand, which they don't.

China should have focused more on this a long time ago. For a long time development of a native Chinese tools and materials sector was kept at drip feed just to provide a hedge if foreign supplies dried out. And, I told so before this whole deal blew up and the sanctions hit.
China needs at least two different suppliers for every single required product on the list, and they need to be invested in. The government needs to step up, but private companies in China need to step up as well. This will likely require pulling people from other sectors of the Chinese scientific community until all the issues are solved and new people are trained for this sector. Meanwhile it is not like the West is standing still. Just recently there were news with regards to improved mask pellicles for EUV and work proceeding on higher intensity light sources and improved numerical aperture lenses for EUV. China needs to have multiple overlapping working teams on all these fronts and that is not going to be cheap and it is not going to happen fast.
 

KenC

Junior Member
Registered Member
There are just too many things to do. It is not just lithography, although that is the most visible component. You need all the tools, and materials including chemicals and masks, then you need to integrate everything. It is one thing to produce a working sample, quite another to achieve mass production with the quality required for these products. Then, even if you have everything, you need to make a production line with all these components and get it working smoothly and effectively enough to compete with existing solution that have been in the market for at least a decade, from vendors who have been in the market for four decades. This was never going to happen in 2 years. It takes 2 years just to build a fab and that is assuming you already had all the components on hand, which they don't.

China should have focused more on this a long time ago. For a long time development of a native Chinese tools and materials sector was kept at drip feed just to provide a hedge if foreign supplies dried out. And, I told so before this whole deal blew up and the sanctions hit.
China needs at least two different suppliers for every single required product on the list, and they need to be invested in. The government needs to step up, but private companies in China need to step up as well. This will likely require pulling people from other sectors of the Chinese scientific community until all the issues are solved and new people are trained for this sector. Meanwhile it is not like the West is standing still. Just recently there were news with regards to improved mask pellicles for EUV and work proceeding on higher intensity light sources and improved numerical aperture lenses for EUV. China needs to have multiple overlapping working teams on all these fronts and that is not going to be cheap and it is not going to happen fast.
Agree, China should have been better prepared. In a way, the Chinese government is too naive, depending on the good nature of the Western world. The same situation also applies to HK and Xinjiang. China allowed the use of seditious text books for years, before realising they were dealing with double dealing subversive elements in such places.
 

Oldschool

Junior Member
Registered Member
There are just too many things to do. It is not just lithography, although that is the most visible component. You need all the tools, and materials including chemicals and masks, then you need to integrate everything. It is one thing to produce a working sample, quite another to achieve mass production with the quality required for these products. Then, even if you have everything, you need to make a production line with all these components and get it working smoothly and effectively enough to compete with existing solution that have been in the market for at least a decade, from vendors who have been in the market for four decades. This was never going to happen in 2 years. It takes 2 years just to build a fab and that is assuming you already had all the components on hand, which they don't.

China should have focused more on this a long time ago. For a long time development of a native Chinese tools and materials sector was kept at drip feed just to provide a hedge if foreign supplies dried out. And, I told so before this whole deal blew up and the sanctions hit.
China needs at least two different suppliers for every single required product on the list, and they need to be invested in. The government needs to step up, but private companies in China need to step up as well. This will likely require pulling people from other sectors of the Chinese scientific community until all the issues are solved and new people are trained for this sector. Meanwhile it is not like the West is standing still. Just recently there were news with regards to improved mask pellicles for EUV and work proceeding on higher intensity light sources and improved numerical aperture lenses for EUV. China needs to have multiple overlapping working teams on all these fronts and that is not going to be cheap and it is not going to happen fast.
for 28nm, if the DUV lithography works, then its 95% self sufficient. They have almost all other equipments and materials.
The only country ever accomplished that feat.
 
Status
Not open for further replies.
Top