A unified standard and interfaces prevents a lot of R&D be wasted on inventing hundreds of new wheel forms.
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
A quantum encrypted communication device developed by the domestic private sector was shown at the "2022 National Cyber Security Awareness Week Cyber Security Expo", it would be quite interesting to see what type of protocol this device use to route and communicate through each, and what type of military application it could apply to
antiterror13
Brigadier
it is quite a big deal for super secure communication for military, it really make the bully worried
This is a very critical point and maybe one of the reasons why NVIDIA until today had still 90% market share.
All this new AI startups advertise their chip raw computing power with very high and competitive TOPS (thousand of billions operations per sec), but beyond the raw power there is the software layer, and is not a small detail.
First version of NVIDIA's CUDA software layer for GPU was released on June 23, 2007, 15 years ago! And it has all but grown hugely in size and complexity by then.
It is not easy for a startup to replicate this gigantic effort in just few years, but without proper software layer, the bare hardware is just like a horse without bridles, it can be very powerful but you can't ride it.
CUDA is proprietary software by NVIDIA, the most natural and smart move Chinese AI firms can do is to cooperate in writing a common software layer, even better if it is open source.
Huawei has the necessary push in terms of financial possibilities and weight to "convince" them.
antiterror13
Brigadier
Any idea how good this GPU is compared to let's say nVidia A100?
1-Like most new technologies there is a hype phase and a practical phase, like for example 3D printing where every single news pundit was saying that the technology will doom China manufacturing, are 3D printer useless? No, they have their use and niche markets that will grow with time. Quantum computing and quantum technology in general is the real deal, quantum phenomena is no just theorical but is also experimentally real and just because we are in the earlier stage of developing that doesn't mean that this tech doesn't have huge potential.
2- "in other words, American spy organizations are willing to honestly review the pros and cons of technology and save money for the country, while Chinese academic institutions rely on deceiving the country and society for personal gain)."
That is patently false, there is not bigger spender in risky crap than the Americans, but that is how a major power keeps its tech advantages
3- This thread is for semiconductors (Equipment, EDA, ICs, materials and so on), I think this type of discussion on quantum technologies belong to the China Science and Technology thread.
2- "in other words, American spy organizations are willing to honestly review the pros and cons of technology and save money for the country, while Chinese academic institutions rely on deceiving the country and society for personal gain)."
That is patently false, there is not bigger spender in risky crap than the Americans, but that is how a major power keeps its tech advantages
3- This thread is for semiconductors (Equipment, EDA, ICs, materials and so on), I think this type of discussion on quantum technologies belong to the China Science and Technology thread.
National standard might be a good thing or it might not. I think more likely, one domestic player in the AI space will be the big winner and maybe a couple of smaller ones by each of the cloud service providers. The big winner looks like Biren technology right now based on the performance of their chips. I think it is more important that people who are familiar with CUDA can easily use the Chinese one. That way, companies outside of China will also be interested in using Chinese AI chips/servers. After all, Chinese companies will want to sell their GPUs abroad.
I took a look at CUDA just now and it looks very much just like C++. Looks like BirenSupa also has a c++ extension
Huawei is an interesting case actually. I haven't seen them buying into any of these promising chip companies. Inspur/Baidu are buying into/working with both BR100 and Kunlun-2. I speculate Alibaba cloud will eventually purchase BR100 also. Tencent mobile is looking to build its own chips and also invested in Enflame. Alibaba cloud developed their own server chip and also invested in Vastai technology.
Huawei doesn't look to have invested in anyone. Prior to the sanctions, it had a very robust chip making business in server and AI with Hisilicon. In fact, Kunpeng 920 is still being used widely 3 years later. Not as advanced as Alibaba's Yitian-710, but pretty good for something designed 3 years ago. They also developed their own Ascend series of AI chips. So, I can only presume that they are working hard at designing a newer more powerful AI GPU with both 12 nm and 7 nm process with their own fab or with SMIC. In many ways, Huawei has been forced to trail blaze for rest of China's semiconductor business.
They are kind of not really comparable.
They seem to have 3 products here
SG100 - using 7 nm process and should be their flagship, but has not offered performance measurement yet.
VA10 - which has very good computational numbers for a chip that uses 150 w of power.
VE1 - which has a lot of power for just 40 to 60 w of power.
Again, all of them seem very competitive. VA10 can perform more calculation than Kunlun-2 and uses as much power as Kunlun-2.
VE-1 looks like a great platform for something that has more specific power consumption requirement.
To put things in perspective, BR100 uses I think 550 w power. A100 uses 300 to 400 w.
Hot pressing equipment for compound semiconductors
Successfully released! Jiangfeng Electronics has made a major breakthrough in the large-scale hot isostatic pressing equipment!
On September 2, Ningbo Jiangfeng Hot Isostatic Pressing Technology Co., Ltd. ushered in a heavyweight product - super-large high-temperature nickel-based alloy. This product is 4.1 meters high and weighs 3.5 tons. It is the current domestic single piece. The super-large high-temperature nickel-based alloy with the largest hot isostatic pressing process has broken through the limits of such products, and it took nearly 4 months to develop.
After two days of sintering, the product was successfully released, indicating that Jiangfeng Electronics' large-scale hot isostatic pressing equipment can provide support for the development of large-scale superalloys in my country and fill the domestic gap.
At the same time, the oversized HIP equipment can also be used for the production of oversized sputtering targets. Previously, due to the lack of suitable equipment for super-large targets in my country, the multi-stage splicing process was still used. This breakthrough also provides technical support for the company to develop an integrated large-size sputtering target, which will vigorously promote the continuous optimization of target performance.
Jiangfeng Electronics continues to focus on strategic needs, build hard-core technologies, and ensure the security of the industrial chain. The application scenarios of super-sized hot isostatic pressing equipment have been greatly expanded again, which will effectively promote the technological improvement and R&D level of related industries in my country's semiconductor material industry chain, and provide strong support for the development and industrialization of new materials in my country.
- Status