1980Di can be used for 14nm production. SMEE first prototype cant. but i am hearing 14nm capable has been made.
1980Di NA =1.35 14nm capable
2000i NA=1.35 7nm capable
2050i NA=1.35 7nm capable higher throughput.
Last edited:
Chinese semiconductor thread II
- Thread starter vincent
- Start date
1980Di NA =1.35 14nm capable
2000i NA=1.35 7nm capable
2050i NA=1.35 7nm capable higher throughput.
They more or less can control how these components are made because these companies probably collaborate with each other but I was referring how using software and the data from the overlay tool SMEE can make automatically make adjustment to the scanner, like alter the dosage, heat the lens system to account for distortions, adjust the MMA for account for aberrations and many other things to "print" the next layer.
Yes correct, SMEE is an integrator and the previous post of our great professor @tokenanalyst, the component suppliers had come of age BUT the latest SMEE iteration the mythical SSA900a 22nm DUVL may had came to late in the game and maybe shelved as SMIC and other FAB operators may concentrate on the expected arrival of Chinese EUVL, as they need all the technical feedback to make it work to mass produced the 3nm node to stay abreast with TSMC. Maybe the tech develop from SSA900 may found its way to SSA800A as both came from the same generation of machine.
This is the year when the 3rd generation of Xiangshan CPU will get released. Claim to be within 8% of ARM Neoverse N2 performance.
None of the high performance RISC-V chips capable of competing with ARM's Neoverse have taped out and are in actual silicon yet. That I know of. But there are a couple running as simulations. Even after these chips become available in the market, it will take several more years to build the systems with these chips, and improve the software support to be able to truly compete with ARM. But I expect this to happen still this decade. Especially in the datacenters of companies like Alibaba.
Other than the Xiangshan RISC-V core design from China there are from the US the Akeana 5000, Ventana Veyron V2, and others.
Other than the Xiangshan RISC-V core design from China there are from the US the Akeana 5000, Ventana Veyron V2, and others.
Last edited:
The screenshot accusing Xiangshan team of building 'PPT CPU' is just lol. They are quite competent people with good reputation in the field.
Neoverse N2 (basically Cortex-A710, Armv9) isn't a top-performing CPU and matching its performance won't be too surprising for me. My main concern is commercialization. Huawei already builds servers with a server-version of K9000s which beats Neoverse N2 easily. There are also Loongson PC and server offerings and Rockchip latest rumors of a new SoC based on Cortex-A730. The performance and eco-system disadvantages for Xiangshan put it in a very similar situation as Loongson so I'm not sure how it will end up.
It's certainly good news for RISC-V though.
NAURA's patent for "a chamber cleaning method and semiconductor process equipment" was published
Tianyancha shows that Beijing North Huachuang Microelectronics Equipment Co., Ltd.'s patent for "a chamber cleaning method and semiconductor process equipment" has been announced. The application publication date is November 29, 2024, and the application publication number is CN119050039A.
The present application provides a chamber cleaning method and semiconductor process equipment, the method includes obtaining the current state of the baffle when the electrostatic chuck in the process chamber meets the triggering condition of charge cleaning, and controlling the baffle to move out of the process chamber when the baffle meets the removal condition based on the current state of the baffle and the first target state of the process chamber, so as to control the process chamber to perform charge cleaning when the current state of the process chamber reaches the first target state, and the charge cleaning is used to neutralize the residual charge of the electrostatic chuck, so that the baffle can be automatically moved out of the process chamber and the charge cleaning can be automatically performed, which greatly improves the cleaning efficiency of the residual charge in the electrostatic chuck, and does not require manual participation, reducing the labor cost in the residual charge cleaning process. At the same time, by moving the baffle out of the process chamber, the influence of the baffle on the residual charge cleaning effect of the electrostatic chuck can be effectively avoided, and the cleaning effect of the residual charge in the electrostatic chuck when automatically cleaning is performed can be ensured.
I guess they are referring to listed companies so this revenue doesn't include private semiconductor equipment companies.
The total revenue of A-share semiconductor equipment companies is expected to reach 99 billion yuan in 2024, with a growth rate significantly higher than other fields.
According to statistics, in terms of operating income, the total revenue of A-share listed semiconductor equipment companies was 20.09 billion yuan in 2019, 25.18 billion yuan in 2020, a year-on-year increase of 25%, 38.37 billion yuan in 2021, a year-on-year increase of 52%, 57.83 billion yuan in 2022, a year-on-year increase of 51%, and 76.76 billion yuan in 2023, a year-on-year increase of 33%. Jiwei Consulting predicts that the total revenue of A-share listed semiconductor equipment companies in 2024 will be 99 billion yuan, a year-on-year increase of 29%, and the compound growth rate of its operating income from 2019 to 2024 will be 30.5%.
Jingzhida: FT test machine has provided samples to customers for verification
Jingzhida released an announcement stating that the company's research and development and production of semiconductor finished product testing (Final Test, hereinafter referred to as "FT") equipment (hereinafter referred to as "this product" and "FT testing machine") have achieved significant results.
FT testing machines are mainly used to perform functional and electrical parameter tests on packaged chips after they are packaged.
Based on the long-term technical accumulation of Jingzhida and its holding and participating companies, the FT tester developed by Jingzhida will make significant progress in 2024. This product uses a dedicated ASIC chip, which can achieve up to 9Gbps signal output and calibration. The number of system channels is 1.5 times that of similar foreign products. It can be applied to finished chip testing such as DDR4, DDR5, LPDDR4, and LPDDR5. The product has completed internal testing of the company, and the test results have met the design requirements. As of the date of this announcement, Jingzhida has provided customers with prototypes of this product for verification.
Jingzhida said that this product enriches the company's semiconductor test equipment product line, enhances the company's market competitiveness, promotes the company's continuous, healthy and stable development, and helps accelerate the upgrading of local industries and ensure the security of the industrial chain. In the future, the company will continue to increase its technical investment in semiconductor equipment to provide customers with more categories and better quality products.
Western WFE is ~500B RMB in revenue. Assuming private Chinese players are roughly 50% the size of their public counterparts, then China WFE is at 150B RMB, or ~30% of Western WFE in scale. Use a PPP deflator and then Chinese players are at 60% in scale