Unless the companies themselves, whether SME, designers, foundries, etc publish it straight from their own sources everything anyone says, whether its from people closer to the industry or observers is still an educated estimate or guess, although some may be closer to the truth then others. Since this is facts/evidence based forum, without statements from the source of truth (the companies themselves) basically everything we say here is a presentation/debate of best guesses/opinions based on observed details. I think it is fair to ask/challenge someone to provide a reasoning/basis for what they stand behind, and @hvpc and other industry experts such have more then provided their rationale for where they stands. Others are free to disagree with this conclusion, and vice versa. It is unfair to rag/attack on him as if he is downplaying progress or denigrating the work being done. I personally think its great to see an inside perspective, helps me learn and put so many things into perspective and I wouldn't want to lose a perspective just because its not as optimistic. We shouldn't let this turn into a positive-reinforcement circle jerk like some others we won't name so lets judge progress and achievements by the results, such as published statements/product output/deliveries from the industry participants and then we can award a medal to who is closest to the truth.
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
Systematical Investigation of Flicker Noise in 14 nm FinFET Devices towards Stochastic Computing Application
1- State Key Laboratory of Fabrication Technologies for Integrated Circuits, Institute of Microelectronics of the Chinese Academy of Sciences, Beijing 100029, China2-School of Microelectronics, University of Chinese Academy of Sciences (UCAS), Beijing 101408, China
3-School of Microelectronics, University of Science and Technology of China, Hefei 230026, China
4-College of Communication Engineering (College of Microelectronics), Chengdu University of Information Technology, Chengdu 610225, China
Abstract
Stochastic computing (SC) is widely known for its high error tolerance and efficient computing ability of complex functions with remarkably simple logic gates. The noise of electronic devices is widely used to be the entropy source due to its randomness. Compared with thermal noise and random telegraph noise (RTN), flicker noise is favored by researchers because of its high noise density. Meanwhile, unlike using RRAM, PCRAM and other emerging memory devices as the entropy source, using logic devices does not require any additional process steps, which is significant for industrialization. In this work, we systematically and statistically studied the 1/f noise characteristics of 14 nm FinFET, and found that miniaturizing the channel area of the device or lowering the ambient temperature can effectively increase the 1/f noise density of the device. This is of great importance to improve the accuracy of the SC system and simplify the complexity of the stochastic number generator (SNG) circuit. At the same time, these rules of 1/f noise characteristics in FinFET devices can provide good guidance for our device selection in circuit design.
2. Device Fabrication and Experiments
The FinFET devices are fabricated based on the Semiconductor Manufacturing International Corporation (SMIC) 14 nm FinFET CMOS process platform. The Fin width (Tfin) and height (Hfin) were 12 nm and 21 nm, respectively, and the effective device width was approximately given by Nfinger × Nfin × (2 × Hfin + Tfin). So, the channel area of the FinFET device is calculated as follows:
Here, A is the channel area, Weff is the effective device width, and L is the channel length
Downplaying, uplaying, pessimistic, optimistic doesn't matter, asking someone for sources as far I understand is not denigrating them. I think to the contrary providing sources will only enrich this thread. Or at least the information should be distilled in a way that make logic sense without giving too much if the person wants. IDK.
If we are going that route, a insider in Shanghai told me that that Huawei and SMIC are making 7-5nm chips in high volume using a shinny Free Electron Laser EUV machine. Sources? I don't want to stir the pot.
If we are going that route, a insider in Shanghai told me that that Huawei and SMIC are making 7-5nm chips in high volume using a shinny Free Electron Laser EUV machine. Sources? I don't want to stir the pot.
Not to put a damper on the enthusiasm, that article is from August 2022 before the October restrictions.
Well fab2 and 3 is already complete so I assume they already running.
And they are expanding in Wuhan also.
Last edited:
Wow, that is unbelievable. They are using Newsonic N40/N41 BAW filters for 5G. Something i would not have expected at all.
interesting that they mention ODM Huaqin's help here in getting the Chinese RF chips into Galaxy, but obviously the companies themselves still have to be competitive
Keep in mind these Newsonic filters are for N40/41 band, so still doesn't seem to have anything for the 77 to 79 band which occupies the higher frequency range
Looks like SMEC has been fabbing their RF filter for a while. I just still haven't seen the higher band stuff
On this one, from April's Shanghai auto show
so because TI's TDA4 became the first obvious one in the market, it had a lot of domestic market share. But as these other local players like SemiDrive, Horizon, Cambrian & black sesame have come on the market, there is not really much advantage in various TDA4 products
By the way, this applies to other SoC and MCUs also from NXP like S32G and Renesas in V3H
All chips that are not that high end in computation. Should be not too difficult to replace
so if firms like Horizon are willing to work closely with automakers to fully utilize their chips, no wonder they get chosen
More importantly, as a single chip is able to do the central ADAS control in the future, you really only need just 1 SoC to do all the handling. Eventually, you will be able to get all you need for both smart cockpit and ADAS.
And if you can do it for something complicated like EVs, then other things should be easy
I appreciate them even if I don't always agree. I wouldn't take opposition to your comments too seriously. It's an unfortunate but inevitable outcome of any large public forum. You get a lot of commentary regardless of how unqualified it is.
I think b40 and b41 are for n40/n41
- Status