But it looks like a good setup to report back to Liu He, and policy adjustment can be done with the feedback from the supply chain chiefs.
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
@krautmeister bro a more detailed report by Tom's Hardware, the Biren tech company stated that its 7nm GPU is competitive with Nvidia next generation 5NM GPU. WOW!!! that is indeed a huge claim and its out of character for a Chinese company to say that unless it's sure of its product performance. And for me I'm sure that it will be FAB by SMIC, I think the author think that with the ban placed on SMIC, it can't be done but we have to wait, I've been surprises a lot this year and expect more to happen.
Things get cloudy
(Image credit: Biren Technology)
Nvidia and AMD continue to dominate the GPU market, but China-based GPU maker Biren Technology hopes to change that. The company is on track to send its first 7nm GPU design to manufacturing in the third quarter of the year; the company said in an interview (via ). Biren Technology says that its compute GPU will compete with Nvidia's next-gen 5nm compute GPUs (thought to be 'Hopper' and 'Lovelace').
Biren Technology's initial product will be a compute GPU tailored specifically for AI workloads, so it's designed for both training and inference. Biren will tape out the GPU in the third quarter this year (meaning it will send the design to manufacturing), and the chip will make its commercial debut sometime next year. Biren's chip will come fabbed on the 7nm node, but it's unclear whether the developer will use Taiwan-based TSMC or its rival, Korea-based Samsung Foundry.
Given the typical amount of time companies spend on bringing up their new silicon (assuming that it works as planned), it's reasonable to expect Biren Technology to release its compute GPU commercially sometime in Q3 or Q4 2022.
Biren Technology received $170 million in Series A funding in 2020, so it does have at funding. Meanwhile, the development and implementation of a moderately complex FinFET-based 7 nm processor costs around $297.8 million, according to estimates by International Business Strategies (cited by ). A highly complex design can cost as much as $1.5 billion, IBS says. To that end, it remains to be seen how capable Biren's AI chip will be and how complex it will be.
Interestingly, Biren CTO and Chief Architect Mike Hong said in an interview with Chinese media recently that Biren's very first chip's performance will rival Nvidia's upcoming 5nm compute GPUs. Of course, being competitive doesn't mean beating the Nvidia GPU, but rather co-existing on the market. Furthermore, there's no way (short of industrial espionage) that Hong could know what Nvidia's codenamed Hopper or Lovelace architectures are capable of, so we should take his comments with a grain of salt.
Biren Technology was established in 2019 to develop GPUs suitable for AI, high-performance computing (HPC), and graphics rendering, a description on the company's website reads. Biren does not plan to build a universal GPU immediately, so the first product will be designed just for cloud AI workloads (i.e., with a focus on INT8 and FP16 computing). Still, it will take advantage of 'high-end packaging technology' to address demanding applications. Eventually, Biren will move on to HPC applications and then will follow with a GPU that be used for AI, HPC, and graphics.
Biren will be the second Chinese company to use a 7nm node to build a GPU for AI and HPC applications, following Tianshu Zhixin with its .
Chinese Company to Send 7nm GPU to Production Next Quarter
By about 16 hours agoThings get cloudy
(Image credit: Biren Technology)
Nvidia and AMD continue to dominate the GPU market, but China-based GPU maker Biren Technology hopes to change that. The company is on track to send its first 7nm GPU design to manufacturing in the third quarter of the year; the company said in an interview (via ). Biren Technology says that its compute GPU will compete with Nvidia's next-gen 5nm compute GPUs (thought to be 'Hopper' and 'Lovelace').
Biren Technology's initial product will be a compute GPU tailored specifically for AI workloads, so it's designed for both training and inference. Biren will tape out the GPU in the third quarter this year (meaning it will send the design to manufacturing), and the chip will make its commercial debut sometime next year. Biren's chip will come fabbed on the 7nm node, but it's unclear whether the developer will use Taiwan-based TSMC or its rival, Korea-based Samsung Foundry.
Given the typical amount of time companies spend on bringing up their new silicon (assuming that it works as planned), it's reasonable to expect Biren Technology to release its compute GPU commercially sometime in Q3 or Q4 2022.
Biren Technology received $170 million in Series A funding in 2020, so it does have at funding. Meanwhile, the development and implementation of a moderately complex FinFET-based 7 nm processor costs around $297.8 million, according to estimates by International Business Strategies (cited by ). A highly complex design can cost as much as $1.5 billion, IBS says. To that end, it remains to be seen how capable Biren's AI chip will be and how complex it will be.
Interestingly, Biren CTO and Chief Architect Mike Hong said in an interview with Chinese media recently that Biren's very first chip's performance will rival Nvidia's upcoming 5nm compute GPUs. Of course, being competitive doesn't mean beating the Nvidia GPU, but rather co-existing on the market. Furthermore, there's no way (short of industrial espionage) that Hong could know what Nvidia's codenamed Hopper or Lovelace architectures are capable of, so we should take his comments with a grain of salt.
Biren Technology was established in 2019 to develop GPUs suitable for AI, high-performance computing (HPC), and graphics rendering, a description on the company's website reads. Biren does not plan to build a universal GPU immediately, so the first product will be designed just for cloud AI workloads (i.e., with a focus on INT8 and FP16 computing). Still, it will take advantage of 'high-end packaging technology' to address demanding applications. Eventually, Biren will move on to HPC applications and then will follow with a GPU that be used for AI, HPC, and graphics.
Biren will be the second Chinese company to use a 7nm node to build a GPU for AI and HPC applications, following Tianshu Zhixin with its .
antiterror13
Brigadier
I think this is the second GPU company producing a competitive chip at 7nm rivaling Nvidia I forgot the name of the other one but as always China provided us gift that keeps on giving.
from CnTechPost
Chinese firm's 7nm GPU expected to tape out in Q3
June 25, 2021
Chinese chip design company Biren Technology's first 7nm GPU is expected to tape out in the third quarter of this year and will be officially released next year, jiemian.com reported Friday.
Founded in September 2019, Biren Technology set a new funding record in China's chip design sector with a Series A round of 1.1 billion yuan ($170 million) in 2020.
It completed its Series B funding in March this year, with cumulative funding of RMB 4.7billion.
A GPU (Graphics processing unit), also known as a display chip, is a microprocessor that specializes in image graphics-related computing work.
Mike Hong, CTO and chief architect of Biren Technology, said in an interview with Chinese media last week that the company's first 7nm chip, which supports both AI training and reasoning, is progressing well and is expected to be formally taped out this year.
Its performance will rival that of NVIDIA's next-generation GPU, Hong said.
By using high-end packaging technology, Biren Technology's 7nm chip is targeted for use in high-end general intelligent computing.
The chip competes with Nvidia's next-generation 5nm GPU, which is still in the pipeline, Hong said, adding that the company's second chip has also started architecture design.
Biren Technology specializes in general-purpose AI training and reasoning capabilities, stripping away designs such as graphics rendering that are unrelated to AI acceleration and focusing more on how to rationally arrange more computing and storage units on its own chips.
Traditional GPUs mainly do vector operations, but for AI acceleration, matrix operations require less bandwidth for data than vector operations.
Instead of sticking to the traditional vector stream processing architecture, Biren Technology will add other elements such as data stream processing units and near-storage computing architecture to its concept to enable it to handle various data types.
Will it be banned and put in entity list ?
@antiterror13 Sir from @vincent previous post.
vincent said:
markets where established players don’t have patents registered.
Update: Huawei denied this report
By
Posted on June 26, 2021
It’s been for quite a time that is preparing to build its own chipsets and now, the latest report on this matter shown us the sign of this possibility.
reports that Huawei will build its first wafer fabrication manufacturing company in Wuhan, Hubei Province. According to the information, the company is expected to start production in stages from next year.
From the first perspective, this is a piece of big news and Huawei’s preparation to enter the chip-making industry is a big topic of discussion. On the other hand, HiSilicon, Huawei’s chip designing subsidiary have the capability to design advanced chip designs.
Sources reveal that this factory initially used to produce optical communication chips and modules to gain self-sufficiency and keep a good quantity of required chipsets.
Huawei is best known for its telecom equipment dominance and the company sells different types of services and consumer products both commercial and for the industry partners.
Huawei Hisillicon Factory in Wuhan:
Currently, Huawei’s Wuhan-based Research Institute has nearly 10,0000 research and development personnel that mainly researching and developing optical communication equipment, HiSilicon chips, and automotive laser radars.
Some media have stated that HiSilicon is the only company in China that has the ability to develop a coherent optical DSP chipset.
It’s reported that the establishment of Huawei HiSilicon’s plant in Wuhan already began to take shape in early 2019. During that time, a project named “HiSilicon Factory” became a hot topic of discussion.
The news of the HiSilcon fab factory broke out from the issues that occurred after the issue in Huawei’s bond in China. In 2019, it’s reported that Huawei intended to register a medium-term note scale of 20 billion yuan, and the first phase of the proposed issuance is about 3 billion yuan, with a term of 3 years. The raised funds will be used To supplement the working capital of the company’s headquarters and its subsidiaries.
In the proposed plan, Huawei proposed the Wuhan HiSilicon factory project with a total of 1.8 billion yuan investment. Related to this, the website of Wuhan Municipal Bureau of Natural Resource and planning also published a “Huawei Technology Co., Ltd. Huawei Wuhan R&D and Production Project (Phase II) A Plot Planning and Design Plan Adjustment Pre-approval Announcement”.
At that time, this project was first mentioned as a HiSilicon factory but later the news was disappeared from the discussion.
Beginning and the next challenge:
For now, there’s no official information that has been confirmed on this subject but we’ll have to wait to get more insights on this developing story.
However, we can say that Huawei’s new semiconductor adventure should get boosted, as it requires more research to develop advanced chipsets. Also, we definitely want to see the Chinese tech company to develop its own chipset for smartphones and other IoT devices.
But at the moment, we can anticipate and joy on this piece of news.
It wasn't that Huawei is collaborating with ICRD so it should be first fab in Shanghai instead of Wuhan?
Imo, any China company advertising any sort of competitive or leading edge performance is inviting sanctions from the US at this point. As long as bottlenecks exist at any point along the supply chain, from raw materials all the way to fully assembled product, China can be sanctioned to oblivion. This is a 7nm design, so it's very dangerous for them to be pre-announcing any GPU leadership. If this were on the 14nm node, it's 1 thing, but on the 7nm node, China won't have fully domestic 7nm commercial mass production capability for at least 2 more years. It's great news, but I always temper my expectations because of the X factor the US presents.@krautmeister bro a more detailed report by Tom's Hardware, the Biren tech company stated that its 7nm GPU is competitive with Nvidia next generation 5NM GPU. WOW!!! that is indeed a huge claim and its out of character for a Chinese company to say that unless it's sure of its product performance. And for me I'm sure that it will be FAB by SMIC, I think the author think that with the ban placed on SMIC, it can't be done but we have to wait, I've been surprises a lot this year and expect more to happen.
SMIC claimed that they will not work with Huawei due to US sanctions, or that SMIC said they will never work with Huawei because of US sanctions.
Last week in state media, it is reported that SMIC was selling 14nm chips to Huawei a year ago (coinciding with the original SMIC pledges not to).
China's 14nm production lines, none have been fully domesticated yet.
I think the kirin 710a chip was getting printed out by smic
Too many dubious or fake news about breakthrough in Chinese semiconductor these days, many in this thread too. I didn't even bother to point them out. Note a lot of these English news are sourced from very dubious Chinese sources, which are either very unreliable or just click-baits. Apply some judgement.
- Status