This is stupid idea. By the time these light sources come up. Lithography will run out of steam and the industry is close to give up on lithography scaling ( shrink ). Dram will go 3D in few years. Sram will be stacked. Materials and DTCO is the way forward.
Chinese semiconductor thread II
- Thread starter vincent
- Start date
Both the wavelength and energy of the wave generated by Japan's FEL based solution appears to be more than a magnitude below what's needed. They're just now "making the case" for the construction of a more advanced test facility to accomplish that, so who knows when the construction may even begin, let alone getting built?
In comparison, Shanghai's SSRF has already produced a light source with the necessary wavelength and average power. The next research facility will be in XiongAn and should start construction soon. I don't know how far along Shenzhen's FEL based solution is.
Shenzhen FEL is a soft xray fel. Not related to lithography at all.
Last edited:
Any summary?
Also, I'm curious what you guys think of the future of Chinese GPU industry. Is Moore Threads the star of the bunch that's most likely to produce decent consumer GPUs? I wonder when they'll be able to offer midrange (4060Ti?) Nvidia level products. Intel started so late and they're already producing viable products
Sometimes I wonder if the lack of consolidation is a bad thing for Chinese semicon. Maybe it'd be better if there was one entity acquiring companies like Loongson, Phytium, UNISOC, Zhaoxin, Moore Threads to create a new AMD-equivalent that can pool resources
Last edited:
While the performance of their integrated graphics has never been meant to compete with discrete GPUs, I'd still assume it's the same kind of engineering proficiency going into these circuits; giving Intel decades of experience.
So, you don't really have any proof that better SoC actually means better phone performance. I've posted a lot of in depth comparison done by Chinese tech vloggers in the past. In terms of speed and battery life, P70 is better than iPhone 15.
Apple continues to show absolutely 0 innovation in its products. Why should it have premium for providing the absolute worst camera & 5g connectivity amongst flagship phones?
Huawei is having trouble producing enough Ascend-910B because HBM scarcity. H20 just cut prices by 20%.
Next gen Ascend is coming out shortly. So if you are a major Chinese tech companies, why would you base your AI future on Nvidia chips? This is actually a pretty complicated issue and your phrasing here is very simple.
Huawei is putting a lot of effort integration Ascend with PyTorch. On top of that, Mindspore has gained significant market share in the past few quarters. It is easily the 2nd most popular ML library. All that shows Chinese private players are increasingly adopting Ascend chips
btw, this is a tightly moderated thread for many reasons. You have shown yourself to really lack knowledge in this area. I would suggest that you read the 300 pages of this thread before continuing to talk, because you have really not shown much knowledge here
There is a EUV FEL in Shenzhen dedicated to lithography.
FEL will come faster than many people think probably in the next 5 years or less. LPP is like the big inefficient mercury lamps that excimer lasers replaced. Problem with current ASML EUV systems apart from stochastics is power which affect throughput, FEL EUV could increase throughput to the level of ArFi scanners or even higher.